Weekly Finds - Week 46, 2020

This week with the Coca-Cola thought experiment, a video about Google algorithm engineers, user-friendly exit popups, gradually changing your mind, and a content marketing report.

Listen to Weekly Finds

Imagine you were the one to start Coca Cola in 1884. How would you do it? Charlie Munger's Coca Cola thought experiment is an exercise in extrapolative and projective thinking. In short: it's a master class in stretching your mind further than you thought possible.

Thought experiments can help us think things through to completion and avoid costly mistakes. Munger draws from areas like consumer psychology, accounting, forecasting, and marketing. My criticism of this is that you need to make a lot of assumptions, and some come out of the blue. The rationale is not always explained in a way the audience can recreate and reapply it. But it's still a valuable business lesson.

Few companies became more popular than Coca Cola. One of them is Google. A new, hour-long video provides more insights into how the Google ranking algorithm team operates.

Google made sure to point some things out very specifically: that revenue is generated through ads (correct), that spam is a major problem (also correct), and that ads are clearly marked (kind of correct; ads have become much harder to distinguish from organic results). Interestingly, at 34:00, Googler's speak about amping up authority in the search results to fight fake news and hoaxes. That could be related to the bigger algorithm changes we've seen in the search results over the last months.

Spam is not just a matter of search results, but also of the onsite experience. Nobody likes being bombarded with intrusive ads or spammy pop-ups. However, some more intrusive tactics still show decent results, so Clearbit ran an experiment to test whether they could create a user-friendly exit popup.

Clearbit was able to design it user-friendly. First, they only showed the popup to visitors who spent at least 30 seconds on the page and capped it at five times. Second, they personalized the popup with Clearbit Reveal, which increased the chance of someone scheduling a demo because it shows the product in action. The experiment led to a 13.4% conversion rate - stunningly good! One important takeaway: where the popup was shown mattered a lot. Pricing and plans landing pages performed by far the best.

Adding more data points to your arsenal, for example with experiments, gets you closer to the truth. In fact, changing your mind in a 180 isn't a trait of smart forecasters. Instead, the best in their fields change their minds gradually.

A study of more than one million predictions shows that some people are “superforecasters” with significantly higher success rates. They gradually update their beliefs. Sudden changes in mind can be signs of recency bias or the availability heuristic. This can be learned, for example by averaging or baselining, which means not giving outliers and recent events too much weight in your predictions.

Animalz updated some beliefs about content marketing with their Benchmark Report 2020.

It looked at 150M pageviews over 12 months and found that

  • Social, email, and communities can kick start a publication
  • Internal linking strategies must be clearly defined and not left up to chance
  • Most articles contain no internal links

To no one's surprise, most traffic came from organic search and social traffic doesn't really matter anymore. However, it seems smaller blogs get less organic search traffic and more from direct and dark social sources.

Of course, we see a power curve in the distribution of traffic: a few blogs get by far the most traffic. More stats from the report:

  • The median traffic is around 17K page views a month
  • it takes about 9 years to reach 1 million yearly page views
  • mega blogs show patterns of having posts that get 100K pageviews each
  • the median bounce rate is at about 80% and time-on-site at 3 minutes and 15 seconds.