Personalization

Quick programming note: I’m going to skip Tuesday’s newsletter and podcast since it’s a holiday weekend in the U.S., and I also need a breather. If you haven’t, check out the episode with Dotdash Meredith CEO Neil Vogel. If you have ideas for the types of guests you’d like to hear from or ways to improve the show, send me a note: bmorrissey@gmail.com.

The 1980s were a golden era for TV. We didn’t have cable TV. Our choices were limited to channels 3 (NBC), 6 (ABC) and 10 (CBS), along with a few fuzzy VHF channels (17 and 29, maybe 48 if the weather was clear) that required intricate maneuvering of rabbit ears to access an inordinate amount of Hogan’s Heroes, Gilligan’s Island and McHale’s Navy. (Side note: I cannot imagine the pitch meeting when the Hogan’s Heroes creators sold a comedy show set in a German POW camp.) No wonder kids played outside more. TV in the 1980s was a shared experience. When I was 10 in 1983, MASH had its final episode after an epic 11-year run. 106 million people watched on average over 2.5 hours of the show. Thursdays in the mid-80s were owned by NBC, when it trotted out a solid three hours of amazing shows, starting with The Cosby Show, then going into Family Ties, Cheers, Night Court and finally Hill Street Blues. Pretty much everyone watched.

Cable TV fragmented these audiences. I remember finally getting cable in the early 1990s and being amazed I could watch everything from Australian Rules Football to the America’s Cup to the odd late-night Shannon Tweed movie. That fragmentation was a forerunner of the internet, which exploded the number of content choices and did away with the artificial constraints of media distribution. Technologists rejoiced, but media people were understandably skeptical. Programming is both science and art. The internet changed that and left it to individuals to program for themselves. The problem: Infinite content requires a sorting mechanism. Enter the technologists armed with algorithms.

The logic of algorithms atomizes content into SKUs, to be reassembled for each individual based on their preferences, often inferred through behavioral data. Each of us could then have our own Thursday night lineup. In 2010, Yahoo’s then-CEO, Carol Bartz, called this the “Web of One,” in one of the many attempts to define what exactly Yahoo was. The magic of “personalization” was taking humans out of the programming, which seems rather impersonal if you think about it. Humans can be terrible programmers but algorithms are even worse since they lack nuance and empathy. Instead, personalization turned publishers from programmers to content producers, sent into the mines to extract content nuggets to be fed into the personalization supply chain to pop up in sundry feeds based on black-box algorithms.

Once the content bundle exploded, so too was the logic of the advertising embedded within it. At its best, advertising was an integral part of the media product. I don't know how many times I heard the idea that people bought Vogue and Wired as much for the ads as the editorial. The ads were a seamless part of the product, and taking them away would make the product worse. Personalization in advertising really meant extracting data for ad targeting in the name of efficiency. Most humans see this for what it is: somewhere between extractive and exploitative. The only people who call ad targeting personalization work in ad tech.

The high opt-out rates for personalized advertising are a signal. The neat trick of retargeting is a technologist and direct marketer’s dream. It’s also on par with pop-up ads for many people. Algorithmic feeds are theoretically personalized, but more people are recognizing how they’re manipulative in service of an engagement-based business model that incentivizes outrage and titillation. When publishers talk about personalization, it is as a subset of optimization, ie, a math exercise. I met with an email service provider many years ago and hearing how we’d get an incremental lift by replacing my role as the editor who chose the stories to feature and replace me with a personalization algorithm. Naturally, I wasn’t too enthusiastic, but I try to be open minded. What’s more, I was told, the send times should be personalized too. Everyone would receive their own version at their own time. This struck me as ludicrous. It might increase clicks but I don’t see how you can build a strong brand if you outsource curation to a computer. If you don’t stand for something, you’ll fall for anything.

Podcasts are a great example of personalized media. Growing them is incredibly difficult. There is no algorithm to be gamed and few optimization techniques that move the needle. I stopped begging people to leave reviews when I found that it would only boost my ego and wouldn’t increase discoverability. (Ego is still important, so please leave The Rebooting Show a review, only if you like it though.) Scott Galloway wrote about how people have a different connection if they’re podcast listeners. He’s onto something.

When I’m out in public, I can tell which media channel is the catalyst for someone approaching me. The bro-approach (i.e., “Yo Dawg”) is made by fans of our videos. Someone who wants to engage in a sober, analytical discussion about the dynamics of streaming or the harms of social media reads the newsletter or my books. But someone who approaches me as if we’re good friends listens to the podcast. It’s similar to running into an old acquaintance you really liked, and who really liked you … but you never had the chance to become good friends. The goodwill is evident. This now happens several times a week.

The human voice is inherently personal, and few people are good enough writers to convey the full contours of their personality. Social media was theoretically personal, but it became manufactured personas. Nobody’s life is like Instagram, and I hope nobody is as psychotic as they seem on Twitter. Podcasts that aren’t overly produced like The Daily are full of human imperfections. (If you want to come to terms with weird verbal tics you have, do a podcast.) I suspect that’s a feature. Bill Simmons spoke to this realization in a podcast he did with Peter Kafka: “People started mentioning it to me on the street — you know, instead of saying, “Love the column,” they would say, “love the podcast.” I thought that was interesting.”

When I think of this concept of humanistic media, I think personalization needs to be redefined and put at the center of the product, including the monetization because that is an integral part of products. Galloway notes how host reads are so effective. Podcast advertising is still so small because, in this case, effectiveness is contrary to the efficiency of scale. Many times, you can have quality or efficiency, but not both.

Instead of treating personalization as yet another optimization technique, what if it was about serving the person’s needs and respecting them as people? What if publishers ran that kind of test for all their various optimization techniques? Would this be how I’d want to be treated? It’s a funny thing that people end up hiding behind data in spreadsheets rather than do some critical thinking. If you applied this test, no publisher would ever require people to call to cancel their subscription because I feel confident, even without data, saying that nobody wants to run the the saves team gauntlet.

Advertising can be personalized without data or targeting. One of the things I like most about doing The Rebooting is that it’s personal by definition. I want TRB to grow, only I don’t want to lose that to become yet another low-engagement publisher that’s fronting another business line. Last week, I had a coffee with a publishing exec who told me he doesn’t mind the ads I run because they fit and they’re (often) written similarly to the content. The more human the ad, the better it performs. To me, that’s a form of personalization, even if there’s no proprietary “decisioning engine” crunching first-party data.

I get notes about how this approach doesn’t scale. I get that. I also don’t concern myself of that much. Starting from zero, you do all kinds of things that don’t scale. But maybe every product doesn’t need to scale. It depends what you want to build. My suspicion, and from what I see happening, is there is a lane for a new group of publishers that aren’t trying to build the next BuzzFeed and choose instead to be meaningful to a group of people, even if it means there’s no IPO.

House of Kaizen collaborates with publishers worldwide to build more sustainable businesses with best practices and experimentation for better acquisition, retention and lifetime value. Matt Cronin, founding partner at House of Kaizen, offers this insight into why consumer expectations and prior experiences are your true competition:

News cycles, media markets and economic environments change dynamically and there’s nothing you can do about it. But subscriber experience is directly within your control.

Subscribers’ experience with myriad subscription products – the best of which are testing and optimizing aggressively – creating the expectations they have for your subscription product.  

Continuously researching, experimenting and evolving the subscriber journey is the only way to wrestle control over the ever-changing marketplace in order to create more sustainable growth.

Learn more about how House of Kaizen helps top publishers and marketers build long-term relationships with their customers — and get in touch with Matt directly.

Recommendations

BuzzFeed’s shitshow of a goat rodeo of a SPAC is a fitting coda to the scale era. It also likely closes the door for a while on any other digital media company going public, right or wrong. Navigating a downturn for these ad-dependent businesses without a Dotdash-like track record of robust profits is going to be rough. BuzzFeed itself is getting so cheap – it is trading far below its expected revenue – that it too could end up being a takeover candidate rather than a the consolidator role that was the entire premise of the SPAC in the first place.

The crypto winter was one I saw coming. I have to admit the trough of disillusionment has been far deeper than I thought, and the resentment of the sector far stronger. There is an understandable glee in taking down the crypto bros and web3 snake-oil salesmen, no doubt. One important takeaway from the current genre of giving web3/crypto boosters enough rope to hang themselves with is the importance of clear, simple communication. I’m reminded of a seminal moment of dot-comenfreude when Razorfish CEO Jeff Dachis told 60 Minutes the i-shop was in the business of “recontextualizing” businesses. He just needed to say, every company needs a website and we build them, but visionaries tend to go galaxy brain. Some similar dynamics are at play here.

From the time you’re a child, you’re always told the importance of setting goals. These days, this has been rebranded into manifesting. I’ve always had a mixed relationship with goals. Like plans, I find the process of setting them more important than seeing them as a linear path. There are simply too many ups and downs and deviations. Having tunnel vision on a goal is a good way to miss opportunities. Sometimes people become slaves to big goals and end up feeling like failures because the don’t hit them. More often than not, that’s not the case. Computer scientist Kenneth has a good reminder that “greatness cannot be planned.”

Substack, which cut 14% of its staff yesterday, is neck and neck with Semafor with most over-covered media company, but I think too little attention is paid to how it is building network effects. Gergly Orosz writes The Pragmatic Engineer, the top technology newsletter on Substack. He recently hit 100,000 free subscribers and detailed how important Substack’s still-newish recommendation feature has been to its growth. I’ve seen a definite uptick in growth since Substack rolled it out in April, and according to Substack, the tool is now responsible for 30% of new (free) subscriptions to Substack newsletters. (My dashboard indicates 64% of new subscriptions over the past 90 days came from within Substack, including its leaderboard.) That’s an impressive start for a recommendations tool that isn’t an algorithm.

If it wasn’t already obvious, time to turn the page on print for local news. Press-Gazette’s list of the top 25 U.S. newspapers by circulation is sobering. Print sales fell another 12% in the past year, and a major metro paper like the Philadelphia Inquirer has a print audience of just 61,000. That’s impossible to make work. It’s telling that the one newspaper on the list showing robust growth is the Villages Daily Sun, the paper of record for reactionary seniors in golf carts.

Thanks for reading and sharing The Rebooting with others. For those in the U.S., have a good (rare) holiday weekend. As always, send me a note with any feedback: bmorrissey@gmail.com.