"What if Google builds it?" 👀
Diffusing platform risks, segmentation that actually conveys how to better retention, and (re)learning to celebrate failure as you scale.
Welcome to the thirty-first edition of The Baton. A fortnightly newsletter that brings you three, hand-curated pieces of advice drawn from the thoughtful founder-to-founder exchanges and interviews taking place on Relay and the interwebz. So, stay tuned!
In this edition, you’ll find instructive and inspiring pickings from the brains of Streak’s Aleem Mawani, Superhuman’s Rahul Vohra, and Gusto’s Edward Kim.
#1: “What if Google builds it?” — Streak’s founder and CEO, Aleem Mawani, has had a long acquaintance with that question and having built a large, profitable business on top of Gmail, he has also figured how to masterfully address it. (Source: Boost VC)
A big risk and a question we got asked all the time was, ‘what if Gmail just builds this’? That’s platform risk.
This is the area where it was helpful to know the team and the strategy at Google. Which is basically, they’re not interested in vertical solutions. They’re interested in ‘how do I get the next billion users?’
They don’t get to the next billion users by having these specialised, vertical use cases. So we knew it wasn’t really a platform risk for us.
From the outside people thought it was a platform risk….You should definitely know the people on the platform team.
Because people don’t realise, it’s really easy to talk, especially for founders who’re like, ‘how do I know what Google’s thinking?’ Just go talk to them! They’re people. Just like anybody else.
…
The other risk with these platforms is, what if they change their products in a way that’s not aligned with your product. So, one worry about Gmail is like, ‘oh what if they change in a way that there’s no threading, no conversations, and whatever.’ And you build your product with that mindset.
And what derisked this for us was that we know email has been around for thirty years. Kinda hard to like really innovate on the model. There’s a bunch of people trying but like it’s pretty clear that the list of conversations is kind of the model.
#2: “I would stop measuring overall churn” — Superhuman’s founder and CEO, Rahul Vohra, sketches out a 90-day plan for tackling churn and retention (and characteristically upends some industry notions.) (Source: CHURN.FM)
I would immediately do two things. First of all, I would stop measuring churn, overall. Many companies do this. They measure their overall churn and start measuring it just for activated users.
Because you want to separate an activation problem from a long-term value problem. And I’m a big believer in optimizing funnels from the bottom upwards.
In this case, the bottom of the funnel is the long-term value proposition of the product not ‘did the first time user experience happened to be effective?’
And so I would draw a line. At superhuman we use, ‘did you, after two weeks of using the product, send 90% or more of your email from superhuman?’ and measure the churn of that group of people separately to the churn of the people who did not meet that threshold.
So that is the first step.
The second step is then I would start to segment by types of users. People who are paying on their personal card versus their corporate card. And you can do some fun things with VIN numbers to determine whether a card is personal or corporate.
I would segment on the size of the organisation. For people who are paying corporately. Are they a sole proprietor? Are they an SMB Are they a medium sized company? Are they a large enterprise?
I would then start segments and other dimensions. Perhaps it is job title, or perhaps it is venture funded versus a more sustainable, traditional business. And I would start to see which segments are doing well, and which segments are not doing well.
I would also then segment against plan type, it’s relatively well known that annual plans are going to have lower churn, often a percent percentage point lower or more.
Once I have this data, I then start to create recommendations. And the recommendations might be, ‘well, it looks like we have a long-term value problem. So we’re going to go back to the basics, we’re going to work on our product-market fit score.’ Or it might be past a certain point, customers aren’t really churning, we just have a tonne of churn upfront.
We’re going to work on the first-time user experience or it might be the stuff above that, which is we’re just not getting the right customers. Or maybe we are, but we’re over-qualifying customers. And we should actually be saying no to way more customers who sign up. And that will have the effects of increasing overall retention.
Or it could be that customers from enterprise or those who pay on their corporate cards, pay, or rather, they retain way longer than people who are paying personally. And maybe we should develop materials and outreach programmes to get people off their personal cards onto their corporate cards.
…But I would take all of these things in my first 90 days, and then go back to my manager and say, this is what the data says, Here are my recommendations, let’s go.
#3: Knowing when you’re optimizing for a fear of failure and what to do about it — Gusto’s co-founder and head of EPD, Edward Kim, cautions against an inevitable outcome of scaling up. (Source: The Engineering Leadership Podcast)
I think it’s helpful to first understand where the fear of failure comes from. And it comes from actually a very good thing, which is, you’ve actually built something in your company that’s going really really well.
…
And your job now goes from trying to get something to work to just trying to make sure you don’t screw it up.
…
But as you continue to scale and you grow, that thing that you optimized for starts to become a disservice to you and the company. Because what happens is this fear of failure, if you take it too far, it starts to change a lot of things about the business. How fast you’re able to execute, your big swings start turning into smaller swings, they start turning into little bunts if I extend that baseball analogy.
…
One thing that we started noticing is that once OKRs were locked in at the beginning of the quarter, people would really optimize for hitting those OKRs. They didn’t want to report back at the end of the quarter an OKR sheet full of red.
And sometimes, we realized that what we had set at the beginning of the quarter was actually not the right thing for us to be working on and yet everybody still optimized to hit these OKRs. There was a strong resistance to change OKRs midway…
That’s kind of a fear of failure, where people optimize for predictability. That’s like a company that’s really oriented around: ‘We don’t want to mess up. We want everything to be predictable.’
…
OKRs are good. We still use them at Gusto, it’s just a different way of thinking about it. At the end of the quarter we’re emphasizing a lot more on what is the impact that you and the team actually had, we don’t really care that much about how accurately did you hit your KRs that you set at the beginning of the quarter.
…
[Yet] another sign of a fear of failure setting in, is when different teams are just optimizing for different failure modes that are against each other [‘in Gusto’s use case, how do I move money fast vs. how do I reduce the risk of losing money’] and you lose sight of like we’re all in here for, to make the company successful, to serve our customers, and grow as fast as we can.
…
I think the solution is to just kind of culturally change through a variety of mechanisms and tell people that it’s okay to fail, obviously there’s a range of failures but certain things where you want to be more tolerant of them
The most important thing instead is to 1) recover very quickly from those failures and 2) make sure you’re learning. You do something to minimize the impact of those failures in the future. And if you get people the permission to do that then the thinking really changes.
…
We say in our engineering world, we’re prioritizing MTTR over MTBF. MTTR is mean time to recovery which is the average time to recover from those failures, we want to optimize around that over mean time between failures.
In fact, we celebrate when people fail and are able to recover quickly.
Until next time,