The immeasurable downside đâ
The "niggling" unease of otherwise entirely worthwhile decisions, notes on modelling early-stage unit economics, and why you shouldn't learn to "let go."
Welcome to the twenty-ninth 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!Â
In this edition, youâll find instructive and inspiring pickings from the brains of Octopus Deployâs Paul Stovell, Mightyâs (and Mixpanelâs) Suhail Doshi , and Oribiâs Iris Shoor.
#1: The immeasurable downside â Octopus Deployâs founder and CEO, Paul Stovell, observes how most well-reasoned, âhighly-measurableâ changes carry hard-to-fathom, potentially-seismic downsides. (Source: Paul Stovellâs blog)
Thereâs this pattern that emerges in engineering, marketing, and a lot of other things that we do. I suspect somebody has already named it - some law or maxim - but Iâve not seen it written somewhere. I describe it as the âhighly-measurable upside vs. the immeasurable downside.â
It goes like this: thereâs a thing that you try - an experiment or a change to a process - and the metrics tell you itâs good. The positive impact of this is highly measurable. You get instant, fast feedback that the change is worthwhile. Everyone would agree you should keep the change.
But thereâs this niggling fear in the back of your mind that makes you feel uneasy with the change. You worry about some backfire or downside. Often, that downside is extremely hard to measure.
You almost feel silly for worrying about it, since the data for the positive case is so convincing, and thereâs no data to be found for the negative case. In a company thatâs very much geared around KPIâs or OKRâs and measuring success, itâs really hard to even raise an objection.
âŠ
Measurable upside: impressions and clicks. A thousand charts in Google Analytics will tell you to keep it up. Youâll hit your MQL quota and get that bonus.
Immeasurable downside: all the people who /donât/ click your ad, and now have a bad impression of you, the word of mouth youâll miss out on, and the great candidates who wonât apply for you, because you come off a little desperate.
âŠ
Measurable upside: The code is 10x faster.
Immeasurable downside: The code is harder to understand now, and weâre afraid to make any changes to it, so weâre not innovating in that space anymore and itâs rot.
âŠ
As an individual, listen to your gut and speak up if you have concerns about potential downsides that fly in the face of the positive metrics. Do work that youâre proud of first, improve metrics second.
As a company, donât use metrics as the primary way you manage people and their performance. Focus on the quality and fidelity of their work. Use metrics to double-check your assumptions and to celebrate, but donât use them to manage. At least this is our philosophy.
#2: Model unit economics early; âItâll keep you honest and let you test the accuracy of your intuitionâ â Mightyâs (and Mixpanelâs) founder, Suhail Doshi, in his usual, confiding fellow-founder voice, lays out how to measure fundamentals, fend off biases, and how not to live in spreadsheets. (Source: Twitter)
I cannot urge early stage startups enough to make a model of their unit economics. It is so telling what will likely cause great, early pain in your startup in the first 5 years. Itâll keep you honest & let you test the accuracy of your intuition. Buffer for the unknown.
For some it may be depressing but for others itâll make you feel a stronger sense of conviction about clever ways you can improve things. That will help you focus more clearly on whatâs important.
When constructing your unit economics model I highly recommend being conservative: assume churn is higher, estimated costs will be greater, improvements will have a reduced impact, average sales prices will be lower. Use similar companies as a benchmark.
When you make your spreadsheet use granular units (hours, GB, users) when calculating costs/revenue. Then adjust those knobs & play around. It helps build your intuition quickly to see whatâs most impactful & what has diminishing returns. And, yet again, greatly clarifies focus.
You can spend many months living in this spreadsheet while in the midst of operating a company. It can feel like the only way to control something you donât directly contribute to as CEO. So, donât! Pick an area to improve & help fix it with the team. Get out of the spreadsheet!
When youâre done, find someone who will lend a critical eye & understands a business like yours to help you diagnose if youâre either missing something or a number is way off. This will help you build conviction that youâre not ignorant about something or extremely biased.
Donât forget that even after you make this model you still need to turn enough of a profit to pay many employees not factored in, facilities, food, etc. If you raised money, you may need to consider how fast youâre allowed to grow if youâre losing money due to runway constraints.
Death by poor unit economics is common. Often the founders will have wished they had that extra 6-9 mo of runway to run the experiments that couldâve turned it around. You may as well know now & start immediately once youâve proved thereâs demand.
#3: Donât âlearn to let goâ â Oribiâs co-founder and CEO, Iris Shoor, diagnoses what can be acutely inhibitive about the âbest practiceâ of letting go and why, as a founder, she has learned to embrace the opposite. (Source: Iris Shoorâs blog)
I believe the basic assumption regarding âlet goâ is wrong: You donât trust people enough; therefore, you are controlling.
My view is entirely different: No one else possesses the full vision of how all the systemâs parts connect. There is a fundamental difference between âgood workâ and âsuitable for the company at the moment.â
The essence of startups (especially at the early stages) is transforming a vision into a realityâan up and running reality. Itâs hard when the vision is unclear, perforated, or constantly changing.
I feel that my involvement in the copy, product definition, and bug-fixing priorities is not an indication Iâm any better than anyone. Itâs because my role is to connect all the dots to create the exact experience I want our users to have. An experience built from so many components that only someone who has a 360-degree view can handle it.
I admit, there were times when I even underestimated managers who insisted on diving into the small details. Today, I understand that thereâs a reason why the most distinguished CEOs were such âdivers.â Itâs impossible to build a robust and innovative vision without being hands-on.
It doesnât mean that Iâm involved in every single detail at Oribi. Some areas operate perfectly without me, areas that are already established and connected to the vision enough. Others areas are less critical to the vision, and Iâm less needed there.
âŠ
And Whenâs the Right Time to âLet Goâ?
We all have areas of expertise we feel are our forte. These are, almost always, the areas that bring us joy. For me, itâs marketing, the productâs roadmap, and finance. It is often hard to give them up.
The biggest trap I see many executives fall into is holding on to a project or a position where they feel at their best rather than passing it on. This is a different âletting goâ than the one this post is about. It comes from fear of stepping out of oneâs comfort zone or the dread of being replaced. If a team member accurately conveys the vision in a particular area, I let that area go (even if itâs my favorite thing to do in the world).
Another aspect is prioritizing⊠At any given moment, the team and I know what matters the most, both within the micro and the macro. Some features are in higher priority than others; there are times when marketing is crucial, and times when itâs a nice-to-have, there are bugs that must be examined and fixed ASAP, and others that I am not even aware of.
Recently on Relay:
AMAs: With Productboardâs co-founder and CEO, Hubert Palan. Featuring: the non-obvious dimensions of customer segments, sorting through the tangle of personas, behaviours, and ecosystems, why Hubert decided to raise funds, and furthering Steve Blankâs work for scale-ups.
Until next time,