The Krehbiel Group - Something I Learned Yesterday

The vast majority of us are average, and average doesn’t sell There’s a lot of concern in the media business right now. It seems as if the world is coming unglued. Brands are dying, or laying off a lot of people. I’ve heard lots of different sorts of explanations, and while there’s some truth to many of them, I’m going to toss my own into the mix. There are simply too many people providing content, and this is what happens. You’ve heard of the Pareto principle – that 20 percent of your customers account for 80 percent of your revenue; 20 percent of your employees do 80 percent of the work; and 20 percent of authors sell 80 percent of the books. It’s a spooky thing, in at least two ways. First, it continues to apply as you did down, so that if you take that top 20 percent of authors who sell 80 percent of the books, 20 percent of those authors sell 80 percent of those books. And that pattern continues to iterate until you get one or two people who sell most of the books. The second spooky thing is that this principle even applies to planets and ecosystems and hurricanes and such, so it’s not just a quirk of human behavior. The internet has made it trivially easy to become a content creator. Bob Hoffman jokes that he finally met somebody who doesn’t have a podcast. YouTube, TikTok, Wordpress – all that, plus the fact that we all have a movie studio in our pocket. This results in an absolute tsunami of content, and AI is going to make it even worse. I just attended a session the other day where a friend explained how you can create a book with designrr.io in about ten minutes. Supply is close to unlimited, but while demand for internet content is way bigger than I ever expected it to be – people spend a huge amount of time on their phones – it’s still limited. There are only so many hours in a day. It stands to reason that the content space is going to have some casualties. Spotify listeners spent 14.9 million total hours listening to The Joe Rogan Experience in the first month it was available on that service. Those 14.9 million hours were not spent on newspapers or magazines. My take on all this is that if you’re a media company, the answer won’t be found in revenue or business model innovation. Those are good things, but they’re not going to address the fundamental problem, which is that the top 20 percent are going to continue to take the lion’s share, and with the constant addition of new creators, that top 20 percent is going to get better and better and better. You’re competing in an ocean of content. New people are coming on board all the time, and while most of them are average, and no particular threat, some of them are talented, and a small portion of them are super talented. My friend Jimmy Finkelstein recently tried to start a new media business on the premise that people want news from a moderate point of view. My first reaction was “No, people actually want to hear lunatics screaming about the end of the world.” But my second reaction was that’s really not the issue. The issue is talent. And I don’t mean to speak against anyone at The Messenger. I’m sure there were some talented people there. But in every profession you have some people who are so much better than everyone else that it defies understanding. Think of Patrick Mahomes, or Steven King. These guys are on the edges of the distribution curve. So my recommendation to media companies is to find extraordinary talent. Things like paywalls and first-party data and all that stuff will pale in significance to the effect of very talented people.

"Eat your own cooking." "The cobbler's son has no shoes." "Physician, heal thyself." "Take your own medicine." There are a lot of sayings that follow that same theme, and I think they're so popular because it's a fairly universal problem. Everyone is good at giving recommendations, but few of us are good at taking them. Even from ourselves. Jordan Peterson says we should treat ourselves as someone we're responsible for. Or, in other words, we should apply the same sort of critical scrutiny to our own actions that we would apply to someone who came to us for help. How do you apply that in your business? Let’s say someone came to you with a complaint that their email marketing isn’t doing as well as it used to. You’d probably have a list of suggestions for them to consider. But do you apply those to your own email marketing? None of us are expert in everything, and we can’t rely on the occasional plea for help from a colleague to prompt us to reconsider our techniques. We need to be more proactive and thoughtful about it. So here’s a simple method. Pick some activity you’re involved in that you wish you could do better at. Step 1: Forget about what you do – your own operations and limitations. Set all that aside. Keeping your own operations in mind will distort the way you look at the issue. You’ll make excuses, and you’ll limit the things you’re willing to consider. Step 2: Come up with a list of sources for advice on the topic. That could include … YouTube videos White papers Blogs Conversations with your favorite AI tool Groups on Reddit Step 3: Spend some focused time digging into these resources – preferably in a new place. Not in your office. Again, don’t think of your own operations. Imagine you’ve been asked to give a presentation on the topic, or write a white paper. Take notes, and don’t worry too much about organizing them. This is the information-gathering stage. Continue to do that until you feel like you’ve heard everything there is to hear, and it starts to get repetitive. Step 4: Forget about it. Put it aside. Spend a day at the park. Sleep on it. Distract your conscious mind from the problem so your unconscious mind can work on it. Step 5: Once again, get out of your office. You don’t want your thoughts to be encumbered by the routine. Review your notes and organize them into an outline. You’ll be surprised at the extra little insights that will occur to you, that can fill in and expand on your outline. During Step 4, your brain was making connections and working on all the things you gathered in Step 3. You’ll find new ways to organize and synthesize the information. Step 6: Now go back to your office, take your outline and compare it with what you’re doing.

Social media is a minefield. You might be asking for more trouble than you want. This may seem like a topic that’s 15 years late in the coming, but companies still wrestle with these things, so here are some thoughts about social media policies. Imagine you’re a publisher and you’ve discovered that when your editors promote their content on social media, it dramatically improves your web traffic, or downloads of your podcast. You’re considering asking (or even requiring) your editors to post their work on social media. Is that a good idea? I have reservations about it, and some suggestions. Reservation #1 – you’re tying your brand to your employee’s personal lives in a way that isn’t fair to them or to you. You’re going to be faced with situations where employees have controversial views. What are you going to do? Are you going to be some Stalinist dictator and police what they do on social media? Or are you going to have your content side by side with who knows what kind of crazy opinion? Neither option is good. Reservation #2 – you’re ignoring personality differences. If you gather 100 employees in a room and give them a talk about how much social media can help with this or that, 10 of them will be gung ho and champing at the bit, 10 of them will want to quit, and the rest will be somewhere else on the continuum in between. This may be a matter of introversion and extroversion, and you’re not going to change that with motivational talks, training, incentives, threats or anything else. People pretty much are the way they are. You’re going to cause your introverts to look for another job. Reservation #3 – some people will be better at it than others. You can train to some level of basic competency, but often there’s a hidden whatsit that makes one person excel and another just get by. It might create resentments, and it’s not your best strategy. Here are my recommendations – if you’re going to do this at all. Suggestion #1 – Ask your employees to post content exclusively on LinkedIn, and also ask them to keep everything they post on LinkedIn professional. No personal opinions. That’s the way most people envision LinkedIn anyway, so that’s not a big ask. And it solves reservation #1. Suggestion #2 – Have a company account on as many social media platforms as you like, appoint someone to manage those accounts, and if you have people who are eager to post content on social media – beyond LinkedIn – let them suggest a few things, try them out, and if they’re good, give them access to that company account. That addresses reservations 2 and 3. Suggestion #3 – Be sensitive to different levels of talent, comfort, and general opinions on social media. Don’t try to force everyone into a particular box. Let me tell a personal anecdote that might shed light on this. Have you ever been to a conference or an event where some excited person gets up front and tries to rile up the crowd. “Are you excited out there?” and all that kind of stuff. Some people enjoy that. I can’t stand it. If you try to motivate me that way, it’ll backfire. Not everybody wants to be on the cheerleading squad. But those other people – the grumps in the corner, like me – also have things to contribute. If you have a few really talented, gung ho, extroverted social media stars, don’t try to clone them. It’s not going to work. Instead, find ways for other people to provide help and support, to make your stars even more effective. Deploy different talents and personalities appropriately. Don’t try to force everyone into the same Myers Briggs category. (In fact, don’t use Myers Briggs at all.) These personality differences mean that you have to give people different options. Don’t tell everyone “please post this on your LinkedIn account,” because the person writing that text is going to be the ra ra “get fired up” cheerleader, and the introverted editor is going to read it and think, “I’d rather die.” So, to sum up. Have a company social media account on lots of social media platforms. Ask your employees to post on LinkedIn only. And accommodate different personalities and levels of talent.

AI models can do a lot of the hard work for content creators Imagine that I have a website with three main audiences. People who love swimming, running, and hiking. Also imagine that I have a customer data platform, so I can identify individual users and keep tabs on what content they like. Since I’m able to do that, I can put people into audiences – e.g., swim fan, running fan, hiking fan. Of course there might be some overlap between audiences, which is fine. Within each audience, I track what’s popular and what’s trending. But not only that. I can check to see if there’s a difference in content preference between the highly engaged visitor and the drive bys and the casual readers. At this point I might have six collections of content – what the highly engaged fan is reading in each of my three audiences, and what the not-so-engaged are reading. Or watching or listening, right? Obviously I can divide this further – like between video and text content – but let’s leave that aside for now and just focus on the content as content. I’m going to pick the highly engaged swim fan as my example. I grab all that content and pull out the keywords, tags, categories, and so on. I’m not going to restrict myself to the keywords and tags assigned by the editors. I’m going to use natural language processing to scan these articles and come up with another type of categorization. Additional words to add to the mix. At this point I need to go on a slight tangent and talk about how AI views words. A word like “king” is represented in an AI system by a multidimensional vector. Something like [0.2, -0.4, 0.7, …] There might be hundreds of dimensions, and the values are assigned so that words with similar meanings are located close to one another in this multi-dimensional space. But they might be close along one axis and not on another. “King” and “queen” are close in the context of being a ruler, but they’re not close in the context of sex, while “king” and “duke” are close in the context of sex, but a little farther apart in rank. These sorts of vectors allow math, like king - man + woman = queen. AI models don’t understand anything. They just have a complicated mathematical representation of words and phrases that are derived from processing huge amounts of text. Let’s get back to swim fans. Once I have a collection of all the keywords that are popular with my swim fans this week, AI can find similar words and concepts and start to build a pretty interesting model of what sorts of topics swim fans care about right now. What’s next? You could just generate a report for the writers showing what types of concepts are playing well with the swim fans. The benefit of using AI here is that this report isn’t just a matter of keywords, like goggles and freestyle and sample workouts. You can get sentiment analysis, like “fun, light-hearted articles are more popular,” or “challenging articles are more popular.” There’s any number of factors you might want to consider. This is analytics on steroids. Your writers can use this analysis to come up with ideas for next week’s articles. Or we can take this all the way to the scary level and ask AI to write the articles based on the information you’re getting from your data. Realize what we’ve done here. We’ve come up with a technological way to listen to our audience, discern what interests them, and then give them more of what they want. While I was going through all this you might have had the thought that there’s a danger of getting into a spiral, where you go deeper and deeper into a narrower and narrower set of concepts. That’s a possible consequence of this kind of work, so it might be a good idea to build mechanisms to prompt people outside of their stated preferences – at least to some extent. For example, Spotify might get the idea that I only like songs similar to the ones I’ve listened to. That’s usually how recommendations work. But it’s a good idea to toss in the outlier from time to time. Anyway, there you have it. A system like that could revolutionize content creation. Your weekly e-newsletter could express exactly what your audience is interested in.

There’s a scene in The Good, The Bad, and the Ugly where Tuco – he’s the ugly guy – goes into a gun shop, examines several revolvers, takes the parts that he likes from each and puts them together to make a revolver that suits his preferences. That’s composability. The problem with buying software is that you buy 100 percent of it, but often you only use 20 percent. For example, your email software can do marketing automation, but so can your customer data platform. So in one place or another you’re wasting that function. And your customer data platform can do content recommendations, but so can Outbrain. So you buy more functions than you can use. When a company buys a new service for their tech stack, they’re often only adding a small portion of that new service’s capabilities. Imagine that you could get a demo of a software suite and say, “That part doesn’t interest me, but your recommendation engine is great. Can I have just that piece?” It sounds lovely. You’d have modular, interchangeable components, like building blocks. But the only way that could work would be if all the pieces had a standard type of connector, like Lego blocks. All Lego blocks have the same size connectors. Software doesn’t work that way. It would be great if it did, but … it doesn’t. Sometimes by design. Apple doesn’t want to make things compatible with Android and vice versa. So what does it mean when people talk about composable customer data platforms? It means they’ve designed the CDP to be modular, allowing organizations to tailor their customer data infrastructure to their specific needs and objectives. The system is supposed to have interchangeable and interoperable parts or components, so that you can plug in any email service provider, any database, any business intelligence tool, and so on. But it’s not quite like Tuco at the gun shop. You can’t pull this piece from here and this other piece from over there, and assemble your own system. For one thing, all these allegedly modular components aren’t always priced that way. You still have to buy components that you don’t want. Another issue is that even if the CDP were completely composable, so that you could pick and choose a la carte, there’s no accepted standard for how all these pieces are supposed to work together. Think of the industrial revolution. We standardized parts so that one factory could make one part, another factory could make a different part, and somebody else assembled it. That only works because of cooperation between the different factories. They have to agree on specs. In the software world, this becomes “integrations.” You need to build special connectors between different modules. While “composability” sounds great in theory. It sounds like we’re taking the lessons of the industrial revolution and applying them to software. But it doesn’t seem to be working quite yet. There are still too many custom formats and custom layouts, so we’re still stuck with building bespoke tech stacks where there’s duplication of function and a lot of inefficiency.

What good is it if someone signs up for your site with their junk email address? Let’s make things overly simple and divide the “how to monetize your website” question into three options. Free, supported by ads Registration wall Paywall There are a lot more options, but bear with me. Each of these models has challenges. If you choose “free, supported by ads,” people can use ad blockers. They can register with their junk email address. And there are ways to get past paywalls. There’s no perfect solution. But maybe we can make them better. You can detect ad blockers and respond with a polite request to turn it off, or you can simply disallow browsers that have ad blockers. There are a range of options. But I want to focus on registration walls, because I don’t think we pay enough attention to this one. The main benefit of a registration wall is that you capture an email address which you can monetize in lots of different ways. But the user can give you a junk email, or they can give you their good email address and unsubscribe from your messages. Why put up with that? It’s analogous to someone subscribing with a bad credit card. If you’re going to limit access to registered users, why not enforce that and insist that the email associated with that registration is deliverable and shows some engagement? This requires an integration between your registration function and your email service provider. This would be a relatively easy use case in a Customer Data Platform. If you’re not sure how to do it, give me a call.

I don’t like social media and I don’t like so-called “influencers,” so keep that in mind as you listen to this one Martha Marchesi proposes four New Year’s resolutions for B2B marketers. Bo Sacks dropped the article in my inbox yesterday, so … here’s my reaction. Yes, no, yes, no. Or, if you want the details … #1 is make friends with AI. I agree with that. Either start using AI or get left behind. Marchesi recommends Perplexity, and she’s in good company. Troy Young from the People vs. Algorithms podcast also has a lot of good to say about Perplexity. And Troy’s a smart guy. I like ChatGPT and Midjourney, both of which Marchesi also recommends, but I dabble in other services from time to time. I’m not thrilled with DALL-E, but that might be because I haven’t tried very hard to learn it. #2 is put your B2B brand under the influence of influencers. My first reaction to that is very simple. Bud Lite. My second reaction is that I can’t imagine someone I would less like to work with than an “influencer.” But my third reaction is that Marchesi may be right and I’m just being a grump. #3 is add more human touch, which somehow quickly turns into “use short-form video.” Obviously I’m a big fan of short-form video – if my 3 to 4 minute podcast qualifies as “short.” I haven’t pursued the super-short videos because I haven’t found a good way to do that yet. Also, I’m not sure my target market is watching those things to learn about publishing and marketing technology. This is where I get confused. People say, “You need to be on Instagram and TikTok,” and I think, first, “I’d rather die,” but second, to the extent that the people I want to reach are on TikTok, I think they’re doing other things, like watching cats. I don’t think they’re in the mood for serious stuff. #4 is shore up your corporate social responsibility story, to which I say run away as fast as you can. Shoring up on so-called “social responsibility” issues sounds like the best possible way to make half your market hate you. Unfortunately, I really do mean “hate you.” You might pick some topic that you believe any decent, thinking person would agree with, and would help your brand, but it seems to me there are no such topics. We live in strange times where people not only believe a bunch of craziness, but get very upset about it. It’s shocking and disturbing, but that’s the reality. I’m a big believer in the idea that “the only thing necessary for evil to triumph is for good men to do nothing,” but I’d keep that in your private life. There have been a lot of casualties from corporations sticking their noses into social issues. So, to sum up. Yes, make friends with AI. Do not get tied up with influencers. Yes, add a human touch. And steer clear of “social responsibility” stuff, unless it’s really basic stuff like picking up trash on the local highway. That would be fine. Links Four areas every B2B marketer should explore in 2024 https://www.fastcompany.com/91013227/new-years-revolution-four-areas-every-b2b-marketer-should-explore-in-2024

When everybody’s crowding the inbox, try the mailbox. My friend Todd Lebo from Ascend2 sent me a new research study on the use of print in marketing. Yes, that’s right. Print. You might think print is dead, but if you didn’t get the memo, I announced the death of “the death of print” quite a while ago. Print publications and print marketing still have an important role, but of course it’s a smaller role than it used to be. Allegra and Ascend2 partnered on some research on this topic. I’ll provide a link below. “58 percent of marketers,” they say, “expect to increase their use of printed marketing materials in the coming year,” and marketers who have used printed materials in the past year are more likely to have seen an increase in revenue. Brochures and flyers are the most popular printed marketing materials, followed by non-post card direct mail and promotional signs. High printing and production costs are the main barrier to more extensive use of print marketing, but ROI remains high. The main thing holding people back from using print is simple preference. They prefer digital channels – possibly because they think digital marketing is easier to track. That may be a misconception, and I’ll provide a link below on that topic. The report discusses a lot of things about the quality of the printing, which is an interesting differentiator. You can make a printed brochure really stand out with specialty finishes, die cuts, high quality paper, and so on. There’s nothing really analogous on the digital side. Marketers who use direct mail say it improves brand awareness and the company’s image. This research study strengthens the idea – at least in mind – that on both the marketing and the product side, print is a luxury option. Don’t use it as the workhorse. Use it for small lists. For your best customers. Use print to make an impression. Links Impact of Print and Mail in Modern Marketing Strategy https://www.allegramarketingprint.com/impact-of-print-in-modern-marketing-strategy Online advertising: The funny, fuzzy math https://martech.org/online-advertising-the-funny-fuzzy-math/ The death of “the death of print” https://krehbielgroup.com/2023/10/21/im-announcing-the-death-of-the-death-of-print/

In my experience, creating summaries of articles, and creating images to accompany articles are the best use of AI right now The Daily Maverick learned some interesting things while experimenting with ChatGPT version 3. I’ll provide the link below. I don’t believe their specific results are relevant to ChatGPT 4, which has fixed several of the problems with version 3, but they do identify a series of things that your should test if you want to try AI yourself. In the list below, I’ll include the things they tried, plus my own recommendations. For article summaries … * Do you want the summary to include every point in the article, or do you want it to entice the reader to read the full article? * Is the summary accurate? Does it capture nuance and context? * Do you want the summary to stay within a certain character, word, or sentence length? * Does the quality of the summary vary based on the structure of the article – e.g., a pyramid style article might get one kind of summary, while a story with a flowery intro might get another sort altogether. * Should summaries vary in tone or style for different sorts of articles, e.g., news, opinion, or feature stories. * Can the summary duplicate the tone of your publication? And does that matter? Then, of course, are the well-known problems of hallucinations and the simple question whether the summaries are any good. Remember that you may be able to correct problems you identify by adjusting the prompts you use. If you’re going to experiment with something like this, remember Gall’s law – that is, start with something simple that works! And make sure you take the writer’s or editor’s workflow into consideration. It’s no good to build something they won’t use. If you find a way to create good summaries using AI, the next question is how to use them. On my blog I put the summary right up front. Some websites have found that increases engagement with the article, but my reason is simpler. I’m not writing a mystery novel, and if the summary is good enough, so be it. Here are some other ideas. Use the summaries in the mobile app, as a popup, or sidebar. You could even create a completely new interface that just has the summaries. Sometimes that’s all people want. You can try text summaries or summaries expressed in bullet points. In addition to summaries, AI can be used to create translations of your articles. I believe you should label things that are generated by AI, unless it’s fairly obvious. For example, I use AI-generated images on my site, but I think it’s obvious they’re from AI, so I don’t label them as such – although I do have a note on the page to explain how I use AI on the site. Links Daily Maverick experiments with in-house AI solutions while learning from its editors https://www.inma.org/blogs/ideas/post.cfm/daily-maverick-experiments-with-in-house-ai-solutions-while-learning-from-its-editors

One reason we still have print magazines is that digital magazines aren’t that good To some people it’s a mystery why print magazines still exist. Weren’t they supposed to have been displaced by digital — a long time ago? Bo Sacks distributed an article on that topic, which I’ll link below. I’ve written before about the various ways print can be better than digital, and I’ll link to some of those articles, but given the fact that the economics of print magazine publishing are awful and getting worse, the more important question is what can digital publishers learn from print. IOW, if the budget is telling you to get more people to switch to digital, make digital better! In that spirit, I’ll mention five things that I think are better about print publications, and some possible ways to do something analogous in digital. I say “analogous,” by the way, because I’m not advocating for skeuomorphs — where some print feature is ham-handedly duplicated online, like the pretend sound of a page turning. Those things are silly. #1 — A print magazine has ads, but they don’t hover over the text and get in your way the way they do online. You don’t have to x out of them so you can see the article. Ideally, they’re on a page by themself. I.e., there’s a page with an article, and a page with an ad. The ad doesn’t usually crowd out the text. Recommendation: Don’t allow ads to interfere with the reading experience. Design your digital magazine for the reader first, then for the advertiser. #2 — A print magazine doesn’t beep at you while you’re reading. It doesn’t try to distract you. Recommendation: Make it easy for the reader to turn off notifications while reading your digital magazine, and don’t use pop-ups or anything like them. #3 — A magazine is random access. You can start at the back if you like. You can flip pages. You can open to the middle. Recommendation: Focus on easy navigation, create something like “flipping pages,” and make it possible to skip many pages at a time — rather than, for example, having to swipe for every new page. #4 — The ads in the magazine are part of the experience. When I was a kid, my brother got some magazine simply to see all the little ads on the last few pages. You know, ads to mail order sea monkeys, or slides for your microscope, or x-ray glasses. As an adult, it’s nice to see ads that are related to the topic. If it’s a boat magazine, I want to see all the cool new boats. If it’s a hunting magazine, I want to see what’s new in crossbows. That’s part of the appeal. Recommendation: Change the way you think of ads. Some people watch the SuperBowl just to see the commercials. Strive to make ads a welcome part of the e-magazine experience rather than an annoyance. #5 — You can leave a magazine open to a particular page. “Hey honey, check out that article on the coffee table” makes sense because the magazine is laying there, open to page 7. You can also dog ear a magazine, or even cut out a page. Recommendation: Make it easy to remember where you are, to bookmark pages, and to share pages. Print does some things better than digital, and vice versa. But there’s no reason why digital can’t catch up. Everything I’ve said here is just my opinion, and a few of the things I like about magazines. Ask your print audience what they like. And don’t do something cheap and easy so you can say, “Oh yeah, we do that” — the way some resorts will install a tiny little ice-skating rink just so they can say they have one. Think about the print experience, and why people like the things that they like. Don’t try to duplicate it exactly, but think creatively about how to fulfill that need in a digital environment. Also, don’t think of your print subscribers as Neanderthals who can’t get with the program. They have reasons to like print. Learn from them. Links Magazines were supposed to die in the digital age. Why haven’t they? https://theconversation.com/magazines-were-supposed-to-die-in-the-digital-age-why-havent-they-217371 I’m announcing the death of “the death of print” https://krehbielgroup.com/2023/10/21/im-announcing-the-death-of-the-death-of-print/ You don’t have to be a Luddite to see the value in print https://krehbielgroup.com/2023/10/11/you-dont-have-to-be-a-luddite-to-see-the-value-in-print/ We need to improve reader comprehension online https://krehbielgroup.com/2024/01/02/we-need-to-improve-reader-comprehension-online/

In today’s episode I discuss design issues with my friend Tricia Okin. Tricia is a user experience designer who helps organizations and businesses design websites and apps both for the benefit of the company and the user. When I think of design, I’m reminded of something my dad frequently said when I was a boy. He’d say “stop trying to help me.” And that’s exactly how I feel about so many websites and apps. Tricia recommends thinking about “escape hatches,” which is a more polite way of saying “stop trying to help me.” For example, autofill and predictive text are both great functions, but sometimes they get in the way and make things worse. Why not have an easy way to x out of it? Tricia also recommends allowing users to curate their own personalization. Take LinkedIn as an example. You have hundreds of contacts, but if you interact with a few people in a week, LinkedIn will show those people all the time – as if you don’t have any other contacts. It would be nice if the consumer / user could adjust the assumptions behind the recommendations. Users need a way to “skip all the guff” and get back to what they were trying to do – not what the site assumes or expects the user wants to do. You can find Tricia at ... https://www.triciaokin.com/ https://www.linkedin.com/in/triciaokin/

Think of all the things you could do with better information about your customers. I’ve done all my work with Customer Data Platforms in the publishing industry, but every industry needs to understand their customer data better, and there are great opportunities – especially in retail. When I go into a store, I’m always thinking about how that store could reach me more effectively if they had a good CDP strategy. So today I’m going to talk briefly about some CDP use cases for retail stores. The basic purpose of a CDP is to gather customer information from multiple different sources to get a more complete view of the customer – in one place. These sources might include point of sale systems, and other in-store data, e-commerce data, customer service information, social media interactions, loyalty programs, information from supply chain and inventory systems, data from your email system, third-party data sources, and other things. By integrating all these data points, the retailer can get a better sense of the customer and provide a better service. Here are a few practical use cases for retailers. Personalized product recommendations. This can be based on that customer’s purchase history, or on the purchase history of lookalike models. That is, people like that customer. These recommendations can be delivered by email, in app, in print, or as coupons on the customer receipt. Speaking of print, my local grocery store sends me a flyer about what’s on sale. By combining that mail data with in-store purchase data, the grocery store could determine who actually responds to these mailings. They could send fewer mailings – or none – to the people who don’t respond to the mailings, and more to the people who do. They could also do custom printing to send me a flyer for things I want to buy and my neighbor a flyer for things he wants to buy. Once I’m in the store, the retailer could display a QR code with a link to the latest bargains. There are many ways to personalize that experience. The simplest would be by location. But if you can link the recommendation with customer data – for example, with a rewards number, or a login – then you can customize those deals. There’s no point in sending me a discount on dog food because I don’t have a dog. Many of the retailers you visit – especially the big chains – already have this kind of information. When you buy a Hewlett Packard printer at Staples, they send you ads for Hewlett Packard ink cartridges. But smaller retailers don’t always have the IT infrastructure to make this sort of thing work. That raises an important question. How big do you have to be to benefit from a Customer Data Platform? I can’t give a generic answer because it depends on your market, your profit margins, and other things. But if you’re curious, give me a call. That doesn’t cost you anything.

My friend Charles Benaiah takes apart the on-going conflict between The New York Times and OpenAI. Bo Sacks picked it up the other day, and I’ll provide a link below. It pains me to say it – honestly it does – but the New York Times is doing God’s work on this topic. This is, of course, the only topic where it’s possible to say that. Here’s the basic problem. Publishers put their content online for “free” – supported by ads – because that was their path to discovery in the search engines. They didn’t have the foresight to put restrictive terms on access to that content. They should have made it clear that the content was only available under certain terms. I mean – why the heck do we hire lawyers? OpenAI took advantage of the ambiguity. They slurped up New York Times content and used it to create a service to compete with The New York Times. That’s disgusting and bad form and all that, but is it technically illegal? That’s what we need to find out. OpenAI claims this is “fair use” – which is an exception to copyright protection. I’m no lawyer, but I’ve been in and around copyright questions my whole career, and I think this is transparently stupid. That doesn’t mean it won’t win in court. The argument seems to be centering on whether or not AI is quoting copyrighted material verbatim. That is an incredibly short-sighted approach, because if the court rules that AI can’t quote verbatim, they’ll just put in a subroutine to make sure they don’t do that, and the fundamental problem will persist. The fundamental problem being that OpenAI is using copyrighted content in a way that the copyright owner never approved. Unfortunately, they didn’t have the foresight to specifically disclaim this use of their content. There’s a lawsuit in England right now where OpenAI has apparently admitted that it’s impossible to train their pet dragon without eating lots of young people. They said "[l]imiting training data to public domain books and drawings created more than a century ago might yield an interesting experiment, but would not provide AI systems that meet the needs of today's citizens." This is a classic “the ends justify any means” argument. I love ChatGPT. Just last night I had a great conversation with it about recipes to make fortified wines. It’s an amazing service. And the neighborhoods controlled by the mafia were pretty safe. Links New York Times: All is fair in love and AI https://mediamakersmeet.com/new-york-times-all-is-fair-in-love-and-ai/ OpenAI admits it's impossible to train generative AI without copyrighted materials https://www.engadget.com/openai-admits-its-impossible-to-train-generative-ai-without-copyrighted-materials-103311496.html

There are two big questions. Propensity towards what? And cause and effect. Many publishers use AI-driven propensity models to great success. I have a link below that discusses how two news publishers use data to retain subscribers. A typical propensity model for a publisher might discover that those subscribers who view at least three pages per session, at least once per week, are more likely to renew. These kinds of models can drive very important actions, which I’ll discuss below. Such models are built by feeding a lot of customer data into an AI model and asking it to identify the customer behaviors or other data that are correlated with a particular action. Relevant behaviors might include time on site, pages per visit, how many e-newsletters the person signs up for, etc. Other data might include demographics, or location. I have an article below that elaborates on the types of data you might feed into such a model. Not all propensity models are the same. One that’s built to predict who will be a repeat customer at a coffee shop is not going to be the same as one that’s built to predict who will renew their magazine subscription. Even within the world of publishing you’ll need different models for different things, depending, for example, on your revenue model. A publisher with an ad-supported site will be looking for different things than a publisher with a subscription-based site. Here are four ways publishers can use propensity models. Identity people who are likely to churn and put them in a different customer journey. For example, you might have customer service call them, or you might put them in a special email campaign. If you sell leads, you can use a propensity model to identify customers who are more likely to buy your client’s product, and charge a higher price for those leads. Let’s say your propensity model finds out that people who came to your site from Facebook are very unlikely to subscribe. If your major source of revenue is subscription sales, you might not want to bother with Facebook. Or, on the other hand, you could put people who come from Facebook into a different segment, not bother to try to sell them on subscriptions, and try to monetize them with ads. Find actions that are highly correlated with success and encourage those actions. For example, you might find that people who subscribe to at least two e-newsletters are more likely to renew their paid subscription, so you could encourage more sign-ups for e-newsletters. That last one makes an assumption about cause and effect that you need to think about. Just because people who subscribe to at least two e-newsletters are more likely to renew doesn’t mean that if you encourage people to subscribe to more e-newsletters they will be more likely to renew. It might work, but it might not. Consider this. Let’s say there’s a correlation between belonging to a gym and maintaining a healthy weight. That doesn’t mean that if you do some grand promotion to get more people signed up at a gym that they’ll all do better at controlling their weight. You don’t want to confuse correlation with cause, but at the same time, correlations can give you some ideas about the types of things you might want to promote. Let’s say a newspaper discovers that people who read the comics are more likely to renew. Even given what I just said about cause and effect, it would be worthwhile to do an experiment and see if promoting the comics section increases renewals. Links How data is used at news organisations to retain subscribers https://mediamakersmeet.com/how-data-is-used-at-news-organisations-to-retain-subscribers/ Engagement metrics for subscriptions https://krehbielgroup.com/2023/10/25/engagement-metrics-for-subscriptions/

The podcast ecosystem is against you, but there are ways to do it People often ask me how to do a podcast that’s only available to paying subscribers. It’s possible, and I’ll explain a few options below. The main challenge is that the entire podcast ecosystem is built on the Big Tech premise of open content supported by advertising. Apple, Spotify, Google, Stitcher, etc., all assume that model. There are a few models that allow for subscriptions, like Patreon, or the subscription options on some of the platforms I just mentioned. The trouble with those is that the platform gets the subscriber information. This has been the problem for decades. Big Tech prefers free content supported by ads, but they will accommodate subscriptions, so long as the platform controls access. They do this because they know the value is in the customer information. The smart publisher wants to keep the customer information for himself, especially if he offers other subscription content. The subscriber doesn’t want one login for the website and another login to access the podcast. That makes the first option for creating a subscriber-only podcast very obvious. Put your podcast behind the same paywall you use for your other content. You might have to upgrade your server capabilities to manage streaming, but that’s all doable. The problem with this option is discovery. People expect to find you on the major platforms – Apple and so forth. If you listen to ads for podcasts, they’ll say something like “find it wherever you get your podcasts.” But those, as I’ve mentioned, all assume it will be free. Let me quickly mention two other options, and then I’ll give you my recommendation. There’s the cheap and simple option that doesn’t require a paywall. Host the podcast in an obscure directory on your own website. Don’t include it in your site navigation, and don’t allow search engines to index those pages. When a new episode is released, send an email to your paying subscribers with a link to the episode. There are obvious downsides to that option, but it’s worth considering, especially if you just want to test the concept of a subscription podcast. The next option goes along with the paywall model above. Host the podcast on your subscriber-only app. But make sure you include decent functionality. You don’t want to frustrate your subscribers. I get a subscriber-only podcast and I’m always cursing their tech. Perhaps the best option is to have two versions of your podcast: a free version, that goes on all the regular platforms, and a subscriber-only version that you deliver on your password-protected website or app. This gives you the benefit of the ordinary podcast discovery process, but also allows you to keep your best content for your subscribers.

Personalization can be very useful, or very annoying Let’s start with some of the geeky details. In order to personalize, you need to have information about the person. That could be browsing history, demographics from some 3rd party source, purchase history, the e-newsletters the person gets, etc. Technically, a web connection is anonymous. But through the magic of cookies, a website can remember who you are. By the way, this whole thing about the death of cookies refers to third-party cookies. I’m talking about first-party cookies, which are not going away. Call me if you need more information on that. You can manage this kind of data in a customer data platform, or CDP. If you’re curious about CDPs, please give me a call, but it’s not necessary to know about CDPs for today’s topic. If a website has a CDP, or similar technology, the site knows who you are when you come back, even if you logged out after your last visit. Your browser still has a cookie that identifies you. You can delete the cookie, but most people don’t do that. Just because the site doesn’t say “Hello Greg” doesn’t mean it doesn’t know it’s me. That raises an interesting question. Is it wise to expose to the visitor that you know who he is? That could be considered helpful in some cases and creepy in others. A generic rule for that might be that the more the person feels like a member, the more it’s okay to show that you know who they are. Another way to put that might be that it’s more acceptable on a B2B site than on a B2C site. But keep the creepy factor in mind. The other day I read “The Risk of Personalization: do people want and trust it?” by Lars Jensen. It’s a good article with some interesting survey results, and I’ll post a link below. Personalization comes in many different forms, and I don’t think we can make simple rules about whether people like it or not. Spotify giving me music recommendations, or Amazon showing me similar products, are very different things than a news site only showing me articles that fit with my preconceptions. They’re both personalization, but one is helpful and the other isn’t. One of the big things to consider about personalization is whether you’re helping the reader or you’re helping yourself. Not that there’s anything wrong with helping yourself. You’re running a business and you have to make money. But it’s a good question to ask to make sure you’re not going in a bad direction. Personalization can be great for you and for your customer. But there’s a line. It’s not a very bright line. There seem to be a lot of factors to consider. Here are some things to keep in mind as you think about personalization on your site, or in your app. Did you get consent? Does the personalization feel intrusive? Have you provided a clear explanation of how and why you’re personalizing your site? Are you protecting your customer’s data, and have you explained that clearly? Have you chosen a type of personalization that feels like manipulation? Does your reader feel as if he’s falling into an echo chamber, where he only hears one side? Could the personalization be embarrassing if somebody else saw it? Links The Risk of Personalization: do people want and trust it? https://theaudiencers.com/decisions/the-risk-of-personalization-do-people-want-and-trust-it/

Aggregated opt-out data might be hiding the important stuff Many years ago I read a science fiction book about a political organization that recruited people from key demographic groups to wear a special watch that would allow them to see and hear all of a politician’s public statements. The watch also measured the wearer’s reactions and fed that back to the control room, where they could see in real time how people were reacting to different statements. That organization’s candidate had some implant in his brain that allowed some technician sitting in the control room, watching all these monitors, to turn dials one way or another based on this feedback, and change what the candidate was saying. The tech would turn up the empathy dial, or the law and order dial, or whatever was necessary. That’s science fiction, but isn’t that what political parties would do if they could? That kind of real-time data is far more valuable than some after the fact “did you like that speech?” questionnaire, because sometimes it’s just one thing in a 45-minute speech that turns somebody off. You need to know what that one thing is. This morning on LinkedIn I saw a post that was so stupid that I unfollowed that person. I don’t know if that person would care, but assuming he did, wouldn’t it be useful to know which post triggered that action? Let’s say this person posts 10 things in a week. Does a weekly average unfollow rate really help, if 90 percent of the unfollows were in response to a particular post? With more granular data you can isolate the particular topic, or theme, or style that’s making people leave. In a similar way, there are themes that show up in some TV shows that are immediate turn offs for me and my wife. As soon as they go that direction, we stop watching. Wouldn’t it be good for the network to know that? In yet another application, I write books, and some of them are available on Kindle. I want to know a lot more than just whether someone bought the book or liked it. I want to know if they made it all the way through, and if not, where they stopped. That kind of feedback could help me craft better books. In fact, I’d like to know on a page by page basis whether the reader is enjoying the book. Amazon doesn’t give me that information, by the way. Now let’s apply this to your email marketing. If you look at aggregated opt-out numbers – over the course of a month, say – you might miss an important signal. Maybe most of the people who opted out did so because of one particular email. Maybe it was too political, or you used a new kind of image, or a new from address. You need to get granular if you want to find out what’s driving your customers’ actions. Look at opt outs on an email by email basis and not just on an aggregated basis. You can do something similar with your web-based content. There are services that can tell you how far down the page people get before they leave. That sort of detail can help you craft a better product.

Several studies, including one that Bo Sacks sent out the other day, suggest that you comprehend and retain more when you read something on paper. There are a lot of possible reasons for this. One is that you’re less likely to be distracted. If you’re reading on your phone, you’re getting notifications and other annoyances while you’re trying to read. A book or magazine doesn’t do that. Another reason is that you engage more of your senses when you read in paper. There’s the smell and the feel of it. And those senses also provide a different level of context, like where you are on the page, whether it’s a left or a right page, and approximately where the text is in relation to the rest of the document. Memory is often a matter of making connections between things, so the fact that you get more inputs when you’re reading in print probably contributes to better retention and comprehension. How do we apply this to digital reading? As I’ve mentioned before, the economics of print are bad and getting worse, so I’m not arguing for a return to print – although I do think it’s a good idea to offer print to the people who want it. The point I’d like to make is that we can probably beat this reading comprehension deficit. We can do things with our digital publications to not only catch up with print, but to do better than print. The first thing is to stop the distractions. Reading on a mobile website is often a horrible experience. The page is actually designed to disrupt your reading, and not by accident. It’s intentional. That’s how you get more ad revenue. But think about that. Ads and reading comprehension are in tension. It’s like they’re mutually exclusive goals. If I want the reader to understand what he’s reading, I don’t want to distract him with a hundred other things. So how do we reimagine ads so that they don’t disrupt the reading experience? There’s a challenge for you. Let’s consider other ways we can make digital reading more reader-friendly. Make it easy for the reader to turn off notifications while he reads the article. That might only be possible in an app, but I think it’s worth pursuing. Provide text-to-speech options. Some people learn better when they listen. Or even listen and read at the same time. Make it easier to take notes, highlight, and then have access to those notes and highlights later. Provide summaries and options for further information – without sending the reader off the page. Give some context to the words on the screen. That could mean using a unique typeface, including more images in the text, or even connecting the text with various kinds of branding. The idea is to make sure the reader’s brain is collecting more information than just the words. Those are some ideas. I like to hear yours. Publishers like to talk about “retention,” but what they usually mean by that word is “keeping the reader as a customer.” I would like to suggest that the other sense of that word – helping the reader retain the information he reads – will also help with customer retention. If your readers are remembering what they read from your site, that’s gotta help. Links Reading print improves comprehension far more than looking at digital text, say researchers https://www.theguardian.com/books/2023/dec/15/reading-print-improves-comprehension-far-more-than-looking-at-digital-text-say-researchers

It’s so bad that some publishers are repeating the rhetoric of their colonizers. I’m trying to think my way through an article titled “Google Could Become New Money Source For U.S. Journalists, Content Creators.” I’ll provide a link below. I’ve stressed the point that Google started indexing publishers’ copyrighted content – without their permission – on the assumption that this was a means of discovery for the publisher. The tradeoff was that Google ingested your content, but you got traffic. I’ve called it a bad deal, but it wasn’t even a deal. It was a shakedown. Or just plain old theft. As I’ve explained elsewhere, that bargain is over. Google and other tech services have created the assumption that they can simply steal your copyrighted content off your website to train their large language models – so they can replace you. Publishers aren’t thinking strategically. Rather than questioning this whole scheme that they were suckered into, their emphasis now is on compensation. But that’s been the trick all along. The tech giants are happy to give publishers dimes while they make dollars. And it’s happening all over again. The deal in this particular case is that Google will pay Canadian publishers 100 million Canadian dollars to serve news content in its search engine. I like getting a check as much as the next guy, but the strategic imperative right now is not to get paid under the current scenario. The strategic imperative is to change the game. Links Google Could Become New Money Source For U.S. Journalists, Content Creators You have to sign up to get access to this article. https://www.mediapost.com/publications/article/391867/google-could-become-new-money-source-for-us-jour.html?edition=132648

Last week I listened to the recent mini episode of the People v Algorithms podcast in which Troy and Alex discuss AI. (Brian was away somewhere.) Alex mentioned how he used ChatGPT to fine-tune the story for his latest game while he was driving to meet Troy for dinner. I've always used chat GPT on my desktop computer, but after hearing Alex's story I downloaded the app on my phone and had a lovely conversation with Hal about my next fiction book. It was an amazing experience. You should try it. Unfortunately, when I came back to restart the conversation an hour later, Hal had lost the thread, which was quite a mystery to me because when I log in to ChatGPT on my real computer I can see the entire conversation. Also, when I asked ChatGPT how I could memorialize the conversation, it gave me a bunch of dumb answers, like maybe I could record it. So ChatGPT still has a ways to go, but it’s pretty amazing. Just this morning I read unCharles’ latest essay, titled “Peas,” in which he evaluates the value of questions in social media. He ranks things as follows. Likes, then comments, then questions. Questions get the top spot because they foster more interaction. Personally, I hate this monetization of crowds. Social media often seems to boil down to this – if you can do something that will gather a lot of eyeballs – and we know perfectly well what kinds of “somethings” gather eyeballs – you become a star and you get rich. We have prominent celebrities these days whose only real talent is getting attention. That doesn’t sit well with my Celtic and Germanic soul. But questions … that’s another thing. A question tells you a lot about the person asking the question. For example, * What he’s wondering about, and interested in * What’s unclear, which may indicate knowledge or education level * What perspective he’s coming from * His emotional state * His problem-solving approach or method * His needs or motivations In short, there’s a whole world of information you can get from a question. Publishers like to ask questions of their audiences, and that’s a good way to increase engagement. My friend Lev Kaye specializes in that sort of thing, so look him up if you want to get more interaction from your audience. But questions are the next step. You know that the era of search is ending, and that people will expect to be able to ask a question and get an answer. By now you should be building a large language model on your content – or at least thinking about it. Take it to the next step. Don’t only answer questions for your audience. Learn something from those questions. Links ==== People v. Algorithms Podcast https://podcasts.apple.com/us/podcast/people-vs-algorithms/id1642958293 Uncharles https://uncharles.substack.com/ Credspark https://www.credspark.com/

I’ve seen a few articles recently about rebuilding trust in media, which is a good idea. Polls show that trust in media is very low, and in my opinion the lack of trust is very richly deserved. Unfortunately, most of the conversation doesn’t get to the root of the problem, which I’ll try to … in a minute. AI is the big story these days, and there have been some big mistakes around AI recently, so sometimes you might get the impression that if media can just fix the AI problem, that’ll solve the trust problem. That’s not true. The trust problem is way deeper. But just to get the AI issue out of the way, the solution seems fairly simple. If you use generative AI, you should make that clear. I think that’s all that needs to be said. The bigger issue with trust in media is that people think the media is lying to them to push an agenda. The sad fact is … that’s true. Not all the time. Not in every story. But yes, it does happen far too often. I can’t give you advice on the lying part – except to stop lying – but there is one aspect of media bias that seems very easy to fix, and it fits in with a very popular word these days. That word is diversity. If you want to fix your bias, make sure your staff at all levels – your writers, your editors, and your management – has people from different points of view. That will help eliminate blind spots. If you have all Democrats, or all Republicans, or all atheists, or all Catholics, you’re going to be biased. There’s no way around that. The example that sticks in my head is from years ago. As you probably know, there are dueling demonstrations on the national mall over abortion. There’s a pro-life rally and a pro-choice rally. Before the pro-life rally, The Washington Post had a small mention on an inside page. Before the pro-choice rally, they had front-page coverage, maps, advice on parking and how to use the Metro, where to get lunch …. It was incredible. It’s no mystery why there was such a discrepancy. The pro-life position is not represented in the staff or management at the Post. So if I see an article on that topic in the Post, I know where it’s coming from. But here’s the real problem. The Post knows this. They’re not stupid people. Despite whatever they claim for PR purposes, they know perfectly well that they lean left, just as Fox news knows it leans right. That’s not an accident. It’s a strategy. Media has discovered that people want to hear things that confirm their biases, so they pick a set of biases and confirm them. There’s no mystery here. It’s not a problem that can be solved because it’s a direct consequence of the media giving the market what they want. Which is biased news coverage. We just have to come to terms with that.

As content distribution channels multiply, should publishers turn to generic brand advertising? Robbie Kellman Baxter distributed a “Customer Retention Metrics Kit” yesterday. It’s worth looking at as an overview of the terms and concepts you should be thinking about if you run a subscription business. In her post introducing the kit, she said “it’s more important than ever to keep the customers you have – and expand the relationship.” I can’t disagree with that, but what does it mean to “keep a customer” in today’s fractured marketplace? The simple answer is to keep them writing checks, but isn’t it more complicated than that? If someone cancels their subscription to your newsletter but becomes a regular listener to your ad-supported podcast, have you lost the customer? We’d like to reduce this to charts and reports and things, showing how customers migrated from one revenue-generating option to another. But that’s impossible for many of the potential scenarios. In the podcast example I just cited, you usually don’t know who’s downloading your podcast. You probably also don’t make as much revenue from one podcast listener as you make from one newsletter subscriber. But what if this formerly barely interested newsletter reader becomes an advocate for your podcast? Maybe he’s worth even more now. As media is distributed across a broader network, with various formats and things, content creators are going to have to find a way to cope with this. Does that mean they have to be more flexible with their revenue projections? The CFO isn’t going to like that. Robbie’s chart made me realize there are lots of concepts that are getting mushed together into a strange media mix. We have the distinction between audience and customer. “Subscriber” has (somewhat to my horror) come to mean both paid and unpaid. (It irks me that people use “subscriber” for people who sign up for their free e-newsletter, but I have yet to come up with a better word.) Among all these groups we have the freebie-seekers, the bargain hunters, the loyal customers, the brand ambassadors, etc. Even among those who like a brand, there’s going to be some fluidity about how they interact with it. They might freely flow between the free e-newsletter, the YouTube channel, the podcast, the app, the subscription magazine …. You can’t track all that – which drives some people crazy. Everything’s about tracking and metrics and attribution. So how do you create a sensible business model with all these moving pieces? One approach might be to run each product as its own business. The magazine does one thing and the YouTube channel does something, and you realize people might go back and forth, but you can’t measure that, and if you can’t measure it you can’t do anything about it, so just proceed as if it doesn’t matter. But this isn’t an entirely new problem. Publishers have faced this basic issue before. We didn’t know who was buying our product on the newsstand either. So maybe this fracturing of the trackable parts of our business could lead us back to the idea of generic brand advertising as an important part of the marketing mix for publishers. Links Customer Retention Metrics Kit https://www.linkedin.com/posts/robbiekellmanbaxter_a-comprehensive-kit-of-customer-retention-activity-7140734829800697856-r0U3

A/B testing can be fun, but there are some things to watch out for Sascha Bossen has a good article on A/B testing that I’ll link below. Today I’ll review his four points, and add three of my own. Pay attention to the novelty effect One panel might outperform your control simply because it’s new. To avoid this mistake, first, make sure to watch performance over time. The novelty effect will wear off. Second, check to see if first-time visitors behave differently than returning visitors. Make sure your test groups are statistically significant. Three is 150% more than two. But that’s statistically meaningless. You need a larger sample size before you can make any conclusions. When you do an A/B split test, you end up with some numbers. Variant A got X clicks out of Y view, etc. Before you make any decisions, you need to check to see if the numbers are significant, and there are some good online tools to help with that. Sascha provides links in his article. Only change one KPI per test That doesn’t mean only test one thing. Some people say every test should only test one change on the page. I don’t agree with that. Sometimes there’s a need to try a completely different design. But Sascha’s addressing a different point, which is what action you’re using to measure the test. For example, let’s say you have an article template. You could have several KPIs for that page, such as time on site, clicking through to sign up for an e-newsletter, or posting a comment. Your A/B test should focus on one of those because errors multiply as you increase the number of KPIs. Try usability testing if you can’t do A/B tests Sometimes you don’t have enough traffic to do a reasonable test in a reasonable amount of time. In those cases, Sascha recommends usability testing as a proxy. I’m going to add three more things to Sascha’s list. Abandon tests that take too long to finish. Sascha hints at this, but I want to make it explicit. If a test takes too long to get a significant answer, you really can’t trust that test. It’s like the old “you can’t step into the same river twice” argument. If the test goes on for too long, are you really testing what you think you’re testing, or are you testing changing moods and currents in your market. For example, let’s say I want to sell a bicycle, and I test two product images: one of the bicycle, and one of a beautiful person on the bicycle. If that test goes on for a while with no clear winner – going back and forth between the two – I’d start to wonder about external factors, like maybe the beautiful person I picked looks too much like a particular actor. That leads to my next point. Avoid testing psychosis. Testing can sound so logical and mathematical, but the more you get into testing the more you realize how many variables you can’t control, and how many extraneous things could be affecting your results. What if your landing page suddenly gets an influx of traffic from a completely different source? What if a new version of a popular browser changes the way your page looks? And when we get into custom audiences, it gets even more complicated. Does one audience prefer variant A and another prefer variant B? It goes on and on, and while you should have an eye out for those things, don’t go crazy about it. Which leads to my final point. When in doubt follow Amazon. They have tons of traffic and they’re constantly testing. What they’re doing is probably as close to good as you’re going to find. So the cheapest way to do A/B testing is to avoid it altogether and mimic Amazon wherever you can. Links 4 learnings about A/B testing after 5 years working on Paid Content at ZEIT ONLINE https://theaudiencers.com/operations/4-learnings-about-a-b-testing-after-5-years-working-on-paid-content-at-zeit-online/

Meeting a customer need is a legitimate service. Getting customers addicted to your service is drug dealer behavior. I got to the pool very early this morning because I couldn't get back to sleep, and while I counted my laps I was wondering about AI and the future of publishing and whether I should change careers and start selling bespoke liqueurs. I finished my laps early, so I went to Dunkin Donuts to get a cup of coffee and think about today's podcast. Earlier I'd read an article on the hazards of social media – how they use psychological tricks to keep you on the site, so they can sell more ads and get more information about you – so they can keep you on the site longer and show you more ads. The goal is addiction. As I sipped my coffee I went to mediamakersmeet.com and read an article on how to keep readers “engaged on your site,” which reminded me of the social media article. And then I looked at my coffee. Dunkin Donuts isn't trying to keep me “engaged” with their store. They’re not using Vegas-style tricks to keep me in my seat so they can play more ads. They just sell me my coffee and we’re done. They do want me “engaged” in some ways. Just as a book publisher hopes the reader will enjoy the whole book and come back for another, Dunkin wants me to enjoy my whole coffee, and they want me to come back every morning. That’s unlikely because (1) I’m poor and can’t afford that, and (2) I drink decaf, so addiction isn’t really on the table. It's still a very different model than the social media model. Dunkin is giving me what I want and then expecting me to go about my day. This contrast applies to publishers because engagement is a key metric for subscription success. If people aren’t using your content they won't renew, and renewal is the lifeblood of the subscription model. But there are two kinds of engagement. The good kind is when I'm enjoying the experience and learning every step of the way. Like watching a really good documentary. Another good side to the documentary is that it comes to an end. It doesn’t try to keep me glued to the TV all day long. The bad kind of engagement is what you get on recipe sites, where they tell you everything you don't want to know about the recipe just to keep you on their site so you see more ads. You’re “engaged,” but you hate it. Most readers don't want to be “engaged,” let alone “addicted.” They want to get their answer and move along. Like the Dunkin Donuts coffee. And like the coffee, they might want to come back regularly. “Attention” and “engagement” are the mantras of the Adtech world. “Keep them on the site, mine their personal information and show them as many ads as possible for as long as possible.” It reminds me of cancer. It’s something that grows without any restraint. There are legitimate reasons to want someone to pay attention to and engage with your content. You might want to get into the customer’s workflow and help them solve a problem, or keep them up to date on a regular basis. Done the right way, it shows that you’re delivering a product that fulfills the customer’s need. Bad engagement is when you’re not trying to satisfy a legitimate need. You promote consumption just so you can show more ads, or get more data. Bad engagement has the marks of cancer and addiction. This is yet another way that Ad Tech and Big Tech have distorted the publishing ecosystem. They take a legitimate thing and push it too far. Links Social Media Must Move Beyond a Broken Ad Revenue Business Model https://www.rollingstone.com/culture-council/articles/social-media-must-move-beyond-broken-ad-revenue-business-model-1234919543/ “Keeping readers engaged within a website’s walls”: How AI is revolutionizing publisher search https://mediamakersmeet.com/keeping-readers-engaged-within-a-websites-walls-how-ai-is-revolutionising-publisher-search/

On our current trajectory, AI means even more crap on the Internet Bo Sacks visits my inbox every weeknight with three articles about publishing. If you’re in the publishing business and you’re not signed up, you should be. Link below. Recently I received this article from Chris Black. “Reading Print Magazines Is an Elite Pursuit” Don’t worry, this isn’t going to be a rehash of the print vs. digital thing, but the following sentence about that transition caught my eye. “The focus in publishing shifted to cranking out many stories a day instead of a select group each month.” E.g., in a magazine. It necessarily follows that this change led to a decrease in quality. That makes sense intuitively, but we’ve all seen it with our own eyes. There’s a lot more garbage out there chasing traffic and “reach.” Mostly because of the need to feed the ad monster. We’re facing a new and more momentous challenge to publishing with the advent of AI. Publishers can now go full crapola and crank out even more articles every day. We can drive down the value of content even further, and stretch meager advertising revenue over an even broader landscape of silly nonsense. Or, we can pause and ask what it is we’re trying to do. Is publishing just a widget factory for words? Are we mass producing whatever will help us chase ad revenue? Or does publishing have a higher purpose of explaining and educating? To some extent it doesn’t matter. AI will result in an increase in low-quality content on the internet, and no amount of moaning and complaining is going to stop that. But doesn’t that create an opportunity? If you’ve followed digital publishing, you know that there will be a lot of publishers who’ll simply say “giddyup” and fly down this highway even faster. They’ll chase the algorithm. If that means posting 45 second AI-generated videos on TikTok with “surprised guy face” in the thumbnail, that’s what they’ll do. It’s just what we’ve seen for decades with a faster engine. Will this create a market for a second type of publisher who believes that human talent is able to do better? It can be a mistake to bet on the intelligence of the general public. When faced with a choice between intelligent analysis and some snarky guy who can make the news funny, a lot of people will choose the snarky guy. It reminds me of the story about Adlai Stevenson. Someone allegedly yelled out, “Governor, you have the vote of every thinking person,” to which he allegedly replied, “That’s not enough, madam, we need a majority.” That’s not an endorsement of the politics of Adlai Stevenson, by the way. It’s just a funny quote that illustrates the problem. In the publishing world, we’ve allowed the ad-driven big tech monster to create a marketplace that prefers nonsense. Is that just our fate? Shall we simply eat, drink, and be merry, for tomorrow we all live in the Idiocracy? Or is it possible to change the market? Everybody tells me there’s a “dark web.” Can there be a “not stupid web”? Or, as Chris Black seems to imply, is the “not stupid web” found on a piece of paper? Links Reading Print Magazines Is an Elite Pursuit https://www.gq.com/story/pulling-weeds-with-chris-black-reading-print-magazines-is-an-elite-pursuit Bo Sacks https://www.bosacks.com/