Innovation at Scale, #10: Making Effective Decisions in the Digital Age
The two types of decisions, made quickly
Welcome to Innovation at Scale, my bi-weekly* newsletter about making innovation work in large organizations.
* formerly weekly, but now moved to bi-weekly for reasons soon to be announced
In Issue 9, I talked about defining, prioritizing, and testing your hypotheses that underpin the success of any new product or service. This is important at the earliest phases, but also true for testing any new design hypotheses. This leads us to a slightly larger topic — how to make effective decisions in the digital age.
What’s different about digital decisionmaking?
The digital era requires different decisionmaking strategies for two reasons: first, the consequences of decisions can be known immediately; and second, more decisions are able to be easily reversed.
Seeing immediate results
In software, you can deploy changes, and immediately see the impact on your success metrics. This is because digital actions are fundamentally measurable in a way that many analog actions aren’t.
If I buy a television advertisement for my product, I might be able to understand viewership numbers for the program itself, but I won’t know who was watching my ad and who got up to make a drink during the ad break. I definitely won’t know if anyone decided “Yes, I will buy that next time I’m at the store,” and barring exceptional circumstances I won’t even know if someone clicked to my website during the broadcast to make an order based on the advert.
With digital advertising, though, I can measure exactly this. How many people saw my ad and clicked through to the landing page? How many of those made a purchase? What was the average order value, which ads performed best, which user segments were most interested in buying? All of these are immediately knowable.
The same goes for product changes. If I change the home screen of my product, I can see immediate impacts on the engagement of my users. How did this impact time per visit? How many users reached the key pages? Can I gather survey data on-site to measure the users’ subjective experience?
While you should wait for statistical significance before making decisions, you can immediately demonstrate the impact of these changes because it is immediately measurable.
Another difference in digital decisions is that, unlike analog decisions, digital decisions can often be easily reversed.
Deploying software means making changes to a code base, but it usually does not mean making irrevocable changes. The former code is still preserved and could be redeployed if there are major problems with the new version.
There are still analog decisions being made, and often times these are much harder to reverse - changes in supply chains, or physical transportation networks, or certain complex technical integrations. We need a way to distinguish between reversible and irreversible decisions.
The two types of decisions
Luckily, Jeff Bezos has done this work for us already! In a 2016 shareholder letter, he clarified a key Amazon principle of making decisions by distinguishing between Type 1 and Type 2 decisions:
Some decisions are consequential and irreversible or nearly irreversible – one-way doors – and these decisions must be made methodically, carefully, slowly, with great deliberation and consultation. If you walk through and don’t like what you see on the other side, you can’t get back to where you were before. We can call these Type 1 decisions. But most decisions aren’t like that – they are changeable, reversible – they’re two-way doors. If you’ve made a suboptimal Type 2 decision, you don’t have to live with the consequences for that long. You can reopen the door and go back through. Type 2 decisions can and should be made quickly by high judgment individuals or small groups.
As organizations get larger, there seems to be a tendency to use the heavy-weight Type 1 decision-making process on most decisions, including many Type 2 decisions. The end result of this is slowness, unthoughtful risk aversion, failure to experiment sufficiently, and consequently diminished invention.
To recap the above, Type 2 decisions are the most common, and are reversible once they have been made; as a result, Type 2 decisions can be made quickly, and can be shipped as hypotheses. Once we have the data to confirm our hypotheses, we can decide whether the original Type 2 decision holds or needs to be revisited.
This gives us a heavy advantage in a digital world, where the logistics of rolling back a decision are trivial, and where the advantages of speed are paramount. Fast decisionmaking is the key to success in fast-changing markets, and forms the bulk of the advantage that most startups have over most corporates.
Speed as a Habit
Dave Girouard’s article Speed as a Habit starts with one major premise:
“All else being equal, the fastest company in any market will win.”
The next step of the argument is even more important: the team that is fastest is the team that makes the fastest effective decisions.
Learning from Eric Schmidt at Google, Girouard suggests that decisions should be timeboxed in advance, to cut off protracted debate, and to build in the responsiveness required to decide at the right time rather than attempt to perfect the decision in advance.
“The process of making and remaking decisions wastes an insane amount of time at companies. The key takeaway: WHEN a decision is made is much more important than WHAT decision is made.”
Combine this with Bezos’ notion of Type 2 decisions, and you’ll see that this strategy is very low risk for most decisions, and that we can push aggressively on the timeline to confirming a Type 1 decision. Type 1 decisions can also be made more effective by driving to a specific deadline. It’s a form of setting success criteria in advance: when you get to the end of the allotted time, you need to go with the team’s best guess. This works as long as the timelines have been set realistically, and the level of certainty is adequate for the impact and reversibility of the decision.
Commitment, not consensus
Committing to decisions at a time means the team has to agree on something to get the decision “shipped.” This has another important effect — it often means that teams will need to “disagree and commit” to get to the finish line. Consensus is not a requirement for effective decisions, but teams do need to commit to executing a decision once it has been made. Because we are making decisions based on our best guess, we will soon have the data to know the right next steps. High-velocity decisionmaking means spending less time building consensus, and more time shipping our best guess in order to learn the right answers after the fact based on the actions of our customers.
Decision velocity and Speed to value
In startups as well as large companies, the opportunities to improve the speed of your decisionmaking are the same ones that help you deliver value to customers and the business faster. Faster decisions mean a faster “Build-Measure-Learn” loop, which means the product aligns with customer needs more closely more quickly. Getting to 80% certainty in the boardroom, with the ability to gather data quickly after launch and roll back any changes if they’re unsuccessful, is the best — and the fastest — way to move your business forward faster.
Each week, I'll include links to articles, books, or podcasts related to corporate innovation, that can help you accelerate the knowledge and progress of your teams.
This week’s recommendations are the two pieces I’ve discussed in detail above. Both are worth reading (and re-reading) in depth and form a key part of my personal toolbox for effective decisionmaking. Please do go read Speed As A Habit and the 2016 Amazon shareholder letter in greater depth - skip to the "Invention Machine" section of the shareholder letter.
A programming note
Due to positive developments that are not quite ready to be announced, this newsletter has moved to a biweekly schedule starting with this issue. I’ve been a bit flexible with my definition of a “weekly” publication schedule so far, so please consider this new schedule to be yet another experiment on my part. This is a Type 2 decision — more data coming soon :)