How Instacart’s CEO uses an Assumption Report Card to make great decisions

Explore

How Instacart’s CEO uses an Assumption Report Card to make great decisions
How Instacart’s CEO uses an Assumption Report Card to make great decisions

How I privately grade the assumptions that inform my toughest decisions.

How I privately grade the assumptions that inform my toughest decisions.

Fidji SIMO

In order to improve our decision-making abilities, we need to judge the accuracy of our assumptions. I’ve created an Assumption Report Card to log my own assumptions and regularly grade them. Read on to see how—and

Copy this doc

⁠

to create your own.

⁠

⁠

The life of an executive is a running stream of decisions. While we judge leaders on the outcomes of their decisions - something that might be better to focus on is the quality of their decision making.

Over 10 years as a Facebook executive, and now as CEO of Instacart, all the decisions my team and I made were the result of processing a lot of assumptions on consumer needs, market dynamics, technological changes, and many other factors. Many times, these assumptions are like a script running in the background of decision makers’ brains, even though making them explicit is critical in improving both alignment and decision-making. In fact, I often push my team to make their assumptions explicit when recommending decisions and trade-offs, because so often, disagreements on specific decisions stem from misalignments on core assumptions. This led to a real “aha” moment for me — that in order to improve our decision-making abilities, we need to judge the accuracy of our assumptions.

For example, at Instacart we’ve seen that the pandemic has completely shifted consumer buying and spending habits - particularly for online grocery — and they continue to change at an extremely fast pace. While the keystone investments we’re making in bringing grocery online with our retail partners won’t change, the strategic decisions we make every single day are based on a set of assumptions and key questions that we’ve aligned on. As an example, will speed continue to be an increasingly important expectation for consumers? How fast does delivery really have to get? We need to make the best assumptions based on these questions and bet on these assumptions to guide what we invest in. This not only allows us to keep up with trends, but more importantly, to anticipate them.

Score assumptions, not decisions

My technique flips the script — don’t judge the decision, judge the assumptions that go into the decision. Given perfect facts, decisions are usually easy, but when are the facts ever perfectly laid out? So the real muscle to strengthen and practice is: how good are the assumptions we are making? Think of the concept of “garbage in, garbage out” — if the data you use for analyses is bad, then your conclusions are bad; the same goes for decisions - the wrong assumptions result in the wrong decisions. This Dilbert comic illustrates the point perfectly.

⁠

Based on this idea, I started tracking my assumptions for big strategic decisions, usually along four big axes: assumptions on evolution of consumer needs, technological changes, overall business landscape, and competition. Then I would rate these assumptions on a scale of 1-3 - where 1 totally missed the mark, 2 is in the ballpark but with some nuance, and 3 is right on target.

Look back regularly — your scores will change

I come back every month and update my scores based on the latest data — how did I do? Have the facts changed? What did we properly predict and what were we off on? Given that many of the decisions that we make play out over very long periods of time, this is a practice that I now reflect on regularly to ensure that we are reassessing our assumptions, so that we can drive better outcomes. And while I do this practice privately every month as a forcing function to make sure I’m not missing a big change, I encourage the teams to do it systematically every six months, in line with roadmap planning. It’s a great tool to train people that it’s ok to revisit decisions when assumptions change. Meanwhile, six months is also enough time so that teams don’t constantly reopen decisions if assumptions haven’t meaningfully shifted.

For example, at Facebook, when we were working on connecting celebrities with their audiences, we initially assumed that the best way to do this was through text conversations — with the tried and true format of "Ask Me Anything". But while we were building this text product, several trends started to emerge; in particular, on the technical side, it became possible to stream live video with such low latency that a full dialog between a broadcaster and its audience was possible. With this technological change, it became obvious that our initial goal of connecting celebrities with their audience would be much more easily achieved through live video than through text, which led to the launch of Facebook Live, a very successful product.

The cost of missing these shifts in core assumptions is massive. When we consider the success of the iPhone, we forget the demise of the Blackberry, the dominant device at the time of its inception. Why did the Blackberry fail? A key assumption - that everyone would prefer a keyboard. A decade after the iPhone, RIM finally launched a keyboard-less phone. Far too late. How do we avoid a similar fate for the decisions we make today?

My dad was captain of a boat and taught me that it is much easier to make a lot of small, frequent maneuvers to stay on course, rather than get off course and have to make one big maneuver to get back on track. What this taught me is that in an ever-evolving business landscape, I need to frequently revisit my assumptions, and reflect these revised assumptions into my decisions. Otherwise, I run the risk of missing some key changes, requiring a much bigger, more painful adjustment later on. While that sounds obvious, it is hard to build a habit and practice that allows leaders to intentionally reassess assumptions and hold themselves accountable for their ability to make the right ones.

Learn from your patterns, and adjust them

Over time, you’ll start to see patterns emerge from the types of predictions you make. For example you might realize that you always tend to predict technology changes in the right way, but miss on policy assumptions, which might reveal a need to spend more time understanding your policy landscape. You might find that you tend to systematically underestimate your competition, or vice-versa be too paranoid about it, which will be a helpful pattern to correct for future decisions. You might realize that you overestimate how fast consumer behaviors change, leading you to invest too much in future-looking bets and not enough in your more medium-term portfolio; vice-versa, the opposite pattern might lead you to always be playing catch-up. Over time, you will want to ask yourself:

What are the types of assumptions that you systematically get right? That way, you can more confidently and heavily bet on your judgment on these.

What are the types of assumptions that you systematically get wrong? For these, you will know to take more time to educate yourself on the landscape, ask for more diverse opinions, and pause longer before deciding on a course of action.

For assumptions that you initially got right, then turned out to be wrong longer-term, what changed? Did you react fast enough to the change and adjust your decisions accordingly? How can you more rapidly recognize when core assumptions have been challenged, and require a change in past decisions?

How can you include these reflections in a team postmortem, so the whole team can learn about their own decision-making process?

A toolkit for your
`Assumption Report Card`⁠
⁠

I’ve historically tracked my assumptions in a large spreadsheet. When I saw what was possible with Coda, I asked my friend

Shishir Mehrotra⁠

if perhaps my assumptions could be tracked in a better way. His team helped me build out a new way to log my assumptions and regularly grade them in an

Assumption Report Card⁠

. A few notes on how it works:

List your assumptions for an upcoming decision.

Every month, review your assumptions and assign each a grade on a scale of 1-3 - where 1 is totally missed the mark, 2 is in the ballpark but with some nuance, and 3 is right on target. Every six months, encourage your entire team to go through the same exercise.

Evaluate your assumption scores over time and across many decision to figure out your own patterns and biases and improve your and your team’s decision-making over time.

If you would like to try this technique and need help implementing it, the Coda team has graciously offered to help. Click this button to get assistance:

Connect with a Rituals Architect

⁠

Thanks to the following people for their help and feedback: Arianna Huffington, Nikhyl Singhal, Luc Levesque, Naveen Gavini, Priya Monga, Lyndsey Grubbs, Shishir Mehrotra, Taylor Pipes, Erin Dame, and Justin Hales.

⁠

Ready to get started?

Start by copying this doc, then head to

Assumption Report Card⁠

⁠

Copy this doc

⁠

Want to print your doc?
This is not the way.

Try clicking the ⋯ next to your doc name or using a keyboard shortcut (

CtrlP

) instead.