I picked a good week to go on holiday. The last two weeks have been relatively slow on real news. But two weeks combined still gives us a meaty briefing. Yesterday's essay was a summary of ~80 essays that combined is arguably the book you may want to read to understand marketing at the CMO level. You can read it here.
Tomorrow’s interview is with Brian Watkins CMO of BulletProof360. Some highlights this week: COVID news, Zoom calendars, Ad fraud, re-targeting cannibalization, magazines, honor code incentives, fax machine AI, work as a “calling” and the lost herb of Ancient Rome.
Onto the briefing.
This briefing is sponsored by GoToWebinar:
Save up to 23% on GoToWebinar Annual Plans!
GoToWebinar eliminates the headache and hassle from webinar hosting. No matter your goal or skill level, you’ll quickly see why so many choose GoToWebinar for easy virtual event management, performance data, videos, polls, and email invites and reminders. Try GoToWebinar for free today!
- I wanted to put this at the top. The new strain from the UK is looking like it is 70% more infectious than the old strain. This means that lockdown policies that kept the old strain at bay (R0~1.0) will result in exponential growth in the new strain (R0~1.7). To get the new strain to “flat growth” would require policies that would have pushed the old strain out of circulation (R0~0.6). I don’t see that happening. Which means we may see massive overflowing hospitals across a much broader geographic area. If you have not already, you may want to re-consider lifestyle choices the same way you did last March/April. I know my family has fallen into a “new normal”, but that new normal may not be enough in the coming months. One of the guys I followed last year (here is an example from March 2nd) has a very long post on the implications of this new strain. Also because the vaccine is out there I expect the media will not cover these new deaths at near the same rate. Here is a chart of what has happened in the UK in the last month as the strain took over:
Follow-up / Meta
- Contra-Experience Follow-up: The big news of the last two weeks has been the utter failure of the US (and Germany, France, Canada… everyone it seems except Israel) to actually distribute the vaccine. Megan McArdle at WaPo argues it is because Public Health Bodies are so concerned with “Talking to each other” that they are missing the big picture. Derek Thompson at The Atlantic walks through the evidence that policy makers still don’t understand the basics on how the virus works. Scott Aaronson expresses his shock at how badly run our civilization is (recommended!). Arnold Kling explains why government organizations can be good at providing services but terrible at new initiatives (due to inability to understand the basics of project management)
- Corona beer: Last February when COVID was just getting started I wrote that contra the consensus, Corona beer would do just fine. The WSJ provides an update, and finds I was right. In fact Corona (the beer) gained significant share this year as consumers shifted consolidated purchases to “safe options”. If a trip to the grocery store is dangerous, you don’t want to cause an extra trip by taking a risk on a new beer that fails.
- InfoHazards: Last briefing I shared the new term I discovered called “infohazards” - information that makes you worse off. Last week I came across this description explaining the themes behind the film Eyes Wide Shut. It doesn’t use the term infohazard, but it could have!
- Amazon buys Wondery: Bezos once said he wanted to have a piece of every transaction on the internet (maybe he should have founded Stripe instead?). Podcasts are a relatively small revenue source right now, but could potentially be huge if they were better able to monetize. As I have written about regularly, Spotify is making a play here. With the acquisition of Wondery, combined with Audible and their (rapidly growing) display ad business Amazon is lining things up to be the other player in the space.
- Zoom to do calendar and email: Zoom went after a crowded area (video conferences) and did very well by just producing a better product than the incumbents. Email and (especially) calendar seems like they could be similar problems to solve. That would suggest they have a shot. But the other way to think about it is: These are different products and it is HARD to make a good product in any of these categories. Many have tried. Marketing agencies tend to start by being good at ONE type of marketing. When they expand into other types it is because they already have DISTRIBUTION and it is easy to sell more to the same customers - not because they become especially good at providing those other services. But does Zoom really have unique access to customers? Is there only advantage that they have better UX people (and a cheaper cost base from their Chinese workforce)? This seems like a stretch from a product in search of being a company. Most likely this is a BATNA so they can sell themselves to one of the big players (Google, Microsoft and Salesforce all have motives and cash to roll them up)
- Peloton acquiring Percor: Speaking of leveraging distribution, Peloton is acquiring Percor. Percor is a manufacturer of fitness equipment. The basic argument for the acquisition is that it is a way to ramp up Peloton's production capabilities quickly (The subscription service has been falling behind on orders given the high demand this year). Percor will beef up production capabilities and solve that short term bottleneck. But Percor also has relationships and distribution with fitness studios and hotels. Usually the company with the sales distribution buys additional products they can push through those channels, but here is an example of the reverse: A great product, buying its way into a giant distribution channel.
- Posturing: This is not a political newsletter, but Jonathan Bernstein (Bloomberg) makes a point that we can see examples of posturing these last few weeks from both Democrats and the Republicans. Many congressional leaders who voted against Nancy Pelosi as leader two years ago, have switched to supporting her now that the margin is slimmer (i.e., they were posturing two years ago because they could, but when it mattered they voted with their interests). Meanwhile many GOP leaders are voting to throw out the presidential election. The question is are they posturing since their vote does not matter, or are the GOP voting in support of the election posturing (and would switch sides if they actually had a majority to do it). Related: “Hundreds” (out of hundreds of thousands) of Google employees have formed a labor union - this is unlikely to change anything directly, but with another group attempting to exert veto power over Google decisions, expect (on the margin) fewer decisions to happen.
- Uber ad fraud: A few layers to this story. In February 2020 Kevin Frisch, Uber’s head of acquisition went on the Marketing Today podcast. Frisch shared details of the “largest ever fraud case” - a case between Uber and their performance marketing suppliers. On January 3rd, 2021 (two days ago), Nandini Jammi, a journalist who calls out companies on their ad placements tweeted a thread explaining the story and how her organization was involved as far back as 2017 (for calling Uber out on many of their placements forcing them to take a closer look at their ad spend, which led to the fraud discoveries). That thread hit #1 on Hacker News. Lessons: (1) Don’t trust your agencies blindly - there is a lot of fraud out there, (2) The ad fraud old news, but it still went viral - you never know what will hit, sometimes you just need to get lucky
- Morning Brew: Morning Brew is a very successful newsletter that got initial traction from “doing things that did not scale” - the founders went to college campuses and made presentations before classes started and then sent around old fashioned paper sign-up sheets. Now Austin Reif (co-founder) has shared how they scaled to the next level with two unusual paid channels: (1) They advertised in other newsletters - they found incredible arbitrage opportunities due to under-monetized rates, and (2) They launched into Instagram story ads very early on. Both are examples of figuring out a new channel before the channel becomes “efficient”. Also see my interview with Adam Doppelt who did something similar in the early days of Apple’s ApStore.
- Incrementality: Re-targeting always looks great when you track it with last touch attribution. But most of those customers were going to come back to you anyway. Every time I have tested it’s true incrementality I have found it is on the order of 70=90% cannibalized. Here is a paper that backs up my estimate (it finds 88% cannibalization). Andrew Capland found similar numbers at Wistia (~85% cannibalization). It doesn’t mean it never works - I worked with one portfolio company that was seeing a 100x return on re-targeting spend, so even with 90% cannibalization it was still a 10x ROI channel - but it never works as well as the vendors tell you it is working
- Logos: AdAge summarizes the three logo redesigns of 2020 that produced the most “snark”. The problem isn’t the snark - people (especially on Twitter) will always complain about change - the problem is that logo changes are almost always for the marketers and not helpful for the brand or the consumers.
- Facebook: In private communications “Facebook managers described important targeting data as ‘crap’ and admitted accuracy was ‘abysmal’”. Take this with a big grain of salt. While there are no doubt clueless companies advertising on Facebook, I expect the majority of spend is more effective on Facebook than any other non-search platform, and these employees are disgruntled and know less than the clients
- Magazines: Sixty new national magazines launched in 2020. It will take a LONG time for old media to die. Trends take longer than you think
- Amazon: A dive into Amazon’s white label business which is getting regulatory scrutiny. White label represents ~1% of Amazon’s revenue. First white label product was batteries (and now represent 33% of online battery sales). 72% of sales come from the top four brands: Amazon Basics, Amazon Collection, Amazon Elements and Amazon Essentials, but there are many others. More at the link.
- Farmville: The original viral social game is shutting down. GamesIndustry has a profile on how the metric-driven design of the game has influenced the broader gaming industry. And Mark Pincus (Zinga founder) has a tweet storm on the story of how Farmville came to be and how they managed the business in the early days.
- Food Delivery: We are in the early innings for third-party food delivery. What will the future look like? Steven Horwitz at EconLib shares some features from a local restaurant that I expect will eventually be scaled by compnies like DoorDash: (1) The restaurant times food preparation so it is ready just as the drivers are arriving for pick up - to increase freshness to the consumer, and (2) They accept the ability to order ahead of time (days in advance) for delivery at a specific time, (3) The ability to share a link to an order so multiple people can pay for part of the order separately (great innovation for office lunches and other larger scale events without single payers). Horwitz claims these innovations are possible because the restaurant has moved delivery in-house, but I expect third party services to be BETTER at this than trying to make it work sub-scale at a single restaurant. They just haven’t got to it yet
- West Point: West Point operates under an honor code. The idea of honor code schools (UVA is another) is that they do not need to monitor cheating as closely because all the students sign-up to the code, and if you are caught breaking the code the ramifications are far more severe than normal schools (including expulsion). Two incentive stories: (1) The remote classes and exams driven by COVID made cheating even easier, so 73 students at West Point were caught - even honor codes are subject to incentives; (2) 55 of the cheaters were varsity athletes (including 24 football players). Historically these athletes would be banned from competition, but the school decided to change their standards after they discovered who did the cheating. Now the cheaters will be given the chance to play in the Liberty Bowl before their punishment is determined later this month…
AI, Machine Learning and Data
- Fax Machines: One suggestion for start-up ideas is “look to industries that are still using fax machines”. Another is, “Build the pick axes for the gold miners”. Combine the two and you have an AI that has the ability to read faxes. Currently being used to monitor COVID record keeping, but obviously broader (if temporary) use cases.
- Nudging: Email nudges to get people to fill in FAFSA applications and college enrollment did NOT have any significant effect. Too often email uses last touch attribution when we really need to look at incremental lift. The latter will almost alway give you a lower measured effect
- Automation impact on wages: Old worry- AI would steal jobs. New worry- AI will enhance jobs and poor countries will lose out.
- Amazon Turk: Amazon Turk is used for “blind” psychology experiments all the time. This study looked to see if the subjects knew what was happening. It seems they do. When asked guess what the "true purpose" of the study designed to test whether they would lie, 50-80% guessed correctly. 93% also believed that the experimenter wanted them to “cheat”. Leads to a lot of questions about the validity of these experiments.
- FinTech Role: CMO role at a PE-backed fast growing Fintech company. Budget in the $100MM+ range. Based anywhere in the US, and they will build a team around this person locally. Third-way CMO, with a spike in performance marketing. Email me if interested. Feel free to forward this email if you know someone who could fit.
- Credentials: The latest stimulus package included a s clause promoting The New River Gorge from a National River to a National Park. This is expected to boost visits to the site by more than 20%. Nothing else about the area will change. This is the power of a credential.
- Netflix Culture: Trung Phan does a nice job of summarizing the main takeaway from the famous Netflix culture document in ten-slides. The argument is that variance is much higher in technology and creative fields than it is anywhere else (i.e., the 100x engineer, the winner-take-all film industry, etc.). Given that the returns on talent are much higher for Netflix than for say P&G - which leads to the need for Netflix’s unique culture.
- Work as a calling: Hanna Weisman is a PhD candidate who studies “work as a calling”. She shared two papers that show that a feeling of a “calling” is related to how good you are at your job, not how interested you originally were or how much money you are making doing it, “overall, viewing work as "your passion" has a lot to do with retrospective narrative construction.” i.e., we tell the story after the fact.
- The Lost Herb: Environmentalists have been worried we would run out of oil or copper or some resource or another for decades. So far there has not been an example of that happening - except for silphium. Silphium was an extremely valuable herb used in Ancient Rome. Julius Caesar stored it in the Treasury and treated it like gold. It was used as a delicacy (“Roman haute cuisine”), but it was also used medicinally, as an aphrodisiac, as birth control (apparently effectively), as livestock feed (when fed to sheep, “their flesh became delectably tender”), and as perfume. It’s heart-shaped seeds may have been the origin of the stylized heart connected to romance to this day. It also seems like the only natural resource humans fully depleted. The last time the herb was identified was some time in the first century AD. Much more at the link.
Keep it simple and stay safe,