How to Decide
This decision framework got us to 30 million users
Happy (crazy) New Year! It is clear that 2021 does not care if we are still suffering from a 2020 hangover. This is usually a time where we set resolutions and make decisions on how our lives will be different from the year before, but despite events around us, we can make choices and decisions. As I settled into the new year, I went on a (socially distanced) walk with a friend and we shared how we were both the youngest born, leaving us less practiced with decision making and choosing.
Life is a series of choices; startups are a series of decisions.
Making decisions is the only way to make your startup move forward. Progress can’t occur without you deciding to make something happen. The best founders bias towards action. They know that it’s better to decide and be wrong than to not decide at all. The more often you try the more opportunity you have to find momentum,and the better you’ll get at decision-making.
You cannot judge a decision by its outcome alone.[a] Sometimes a call by a coach did not turn out well, but many people given the same knowledge and circumstances would call The outcome can factor in as a data point (with a weighted ranking) when making that decision again. A good decision takes into account factors and weighs them to come to a logical conclusion to eventually produce the best outcome, especially in repeated decisions.[b] If you have a good repeatable process when it comes to decision-making, you can use the outcomes from each decision to keep perfecting your framework.
The process of decision-making is more important than the outcome.
For a founder who was not well-practiced at making decisions, I have several shortcuts for you:
Decide what type of decision you are making
There are two types of decisions: one-way door decisions and two-way door decisions[c]. One-way door decisions are decisions that cannot easily be undone: how to split co-founder equity, fundraising, whether to exit. In life it’s who you marry or whether you have kids. [d] These decisions tend to have a binary answer of yes or no, because the question usually is will you or won’t you. You either have kids or you don’t, you either sell your company or you don’t. These decisions are very costly to reverse if you can reverse them at all. Two-way door decisions are decisions that you can reverse. They are made frequently and tend to have many possible answers. What will you eat today? What types of product features should we release?
Oftentimes founders (and non-founders) mix these two types of decisions. During YC I saw many founders wait until timing and the product were perfect in order to “launch,” treating it like a one-way door decision. The reality is the cost of waiting to launch your product perfectly is far greater than launching until you “feel” ready. The more you wait, the more attempts you lose to get to product market fit. Therefore, you can and should launch your product over and over and over treating it like a two-way door decision. Conversely, one-way door decisions mistaken as two-way door decisions can have major consequences. For founders, I’ve seen this play out when they are ambivalent about fundraising and it’s such a strain on their business. They don’t clearly start or clearly stop, and end up in a muddled mess.
Once you’ve decided what type of decision you are dealing with, here are some other shortcuts to help you make better decisions.
🏗️ Shortcut #1: Set up a Decision-Making Framework
Two-Way Door Decisions
If it is a two-way door decision, such as what features to build for a software product, then you should set up an experimentation framework. The framework should have a clear hypothesis and experiment design. Then ,perform the experiment, collect data, and document your conclusions. In the startup world there is a lot of experimentation to gather data to see what impacts the outcome and decide what to do in the next experiment. Having a standard framework, enables you to answer the same question over and over, tweak your process and get data to inform your future decisions.
After I drew the steps, it seemed to take the similar shape and steps as Framework for Startup Progress from this article. At every step, there should be a system to help you execute several choices at once to get data and learn faster.
An example of this was our system for product design. Product development and design is a perfect example of two-way door decisions. This was our system for making product decisions. It was the same system that got us to 30 million users for our TV shows and community apps:
Set Goals and Success Metrics. We wanted to move the needle in 4 categories: acquisition (growth), retention, engagement or monetization. Initially the goal was focused on growth. But quickly it included retention. If you are a designer, wondering how it fits in design, read Designing for Dollars: How design can impact the bottom line.
Design / Make A Plan. Designers and product managers would meet weekly to discuss business goals and the features to design and build that week. Then we would talk about how we expected the metrics to move and and by how much. We would then create an easy grid that measured the features that we brainstormed and then ranked the features against the following: Level of Impact and Level of Effort. We then would prioritize the ones we thought had a high Level of Impact and the lowest Level of Effort. Finally,we would take a step back and ask “does this make sense”? This pause enabled us all to do a gut check and allow intuition to come in. Intuition should not lead but should be the backstop in these decisions. This enables you to go through a process and separate the process from the outcome.
Execute the Experiment. We would then design the features and work with the product manager to ensure the design was going toward the vision. We would work with engineers on feasibility right after our product meeting which was usually Monday, mid-morning. By Wed morning, stuff was mocked and engineering would be released by Thursday or Friday morning. However conversations with engineering happened right after the product meeting on Mondays. We avoided launching on Friday afternoons because if anything went wrong outside partners weren’t around to help fix it.
Oftentimes founders get really excited about launching the product, but they forget to track the success of failure of their experiment. They don’t put in analytics. Analytics helps us track if the feature was a success or not. Today there are many tools that enable you to track (Google Analytics/Mixpanel), execute and track several experiments at once (Optimizely) and track events (Segment). Before all this we had to create a link tracking tool and manually assign a code to see what links were clicked on and what frequency. It’s so easy now, founders have no excuse.Analyze Results. Every Monday morning, we would have a company meeting and share what we released and how it performed with the whole company. Knowing that there was an all company meeting on Monday kept us accountable to ensure our decision-making and experiments made sense.
At each step we had a system and framework to execute iterations of experiments so that we could focus on impactful product decisions rather than re-designing the process. For two-way door decisions make sure your process is repeatable as you should be using it over and over.
One-Way Door Decisions
One-way door decisions require more deliberation, need less of a repeatable framework, and sometimes you need to seek the advice of others. Ultimately as a founder, to make these decisions, you must dig deeper to what you want. [e]
One question I like to ask with these decisions is around regret.
Will I regret doing this or not doing this more?
For these decisions having a framework guided by goals and values will be incredibly important to bring clarity. [f] Once you have clarity you can focus. And you always have the best chance to progress if you only focus on that one thing.
Clarity brings focus.
Focus brings execution.
Execution brings progress.
Progress brings achievement.
If you want to be a little more sophisticated, you can add a time frame. I fully recognize that it’s difficult to predict how you feel in the future, especially happiness [g], but have found it incredibly clarifying on digging into what you want and how you will build your life or company.
Will I regret doing this or not doing this in 10 minutes?
Will I regret doing this or not doing this in 10 months?
Will I regret doing this or not doing this in 10 years?
🙋♂️Shortcut #2: Hire someone who can complement you on decision making.
The best early hiring decision we made was our first Head of Product. While he had a great pedigree, he is fabulous in two things a) biased towards making decisions and b) putting a framework or process together to come to those decisions. He was the one who set up our framework. He facilitated the questions and arguments and we came to agreement. If the data did not show that our feature release was great, we were able to go back to experimenting with different designs and factors rather than re-work the framework. That is where you know you have a good flexible framework for two-way door decisions. For one-way door decisions, consult with people and learn from their experiences and form a collection. If there is a repeated theme or pattern, you should factor that into your decision making process. This is one reason why investors can be good at seeing trends, because they work with so many founders as a part of their job and can see a collection of experiences at the surface.
✅ Shortcut #3: Set milestone goals and batch cohorts in decision-making
Sometimes the feedback cycle of two-way door decisions is so long, you feel like it’s a one-way door decision--but if you look more closely it’s not. Examples of this are investing, hiring an executive, or whether or not you went to the “right” college. With these decisions you can figure out a shorter way to make progress to your goals. One way of doing this is creating milestone goals. We do this with marriage: first we date, then we get engaged and then we get married. Each milestone has a length of time. If you are choosing between the same kind of choices, you can batch these choices together while running through the milestone goals. In hiring a person you can resume reviews, screener calls, interviews all in cohorts. For investing, YC not only did batch investing but this also enabled them to use a portfolio strategy to hedge their bets and make several simultaneously enabling them to learn quickly. [h]
🚫Shortcut #4: Don’t confuse making a lot of decisions with progress.
You don’t have to make every decision. Sometimes not all decisions are worth making because they aren’t impactful. Other times, if you can make the first decision then the subsequent decisions will become a lot easier to make where you do not need to decide. Make sure each decision you make will impact greatly and directly your end goal, otherwise you are just making a lot of decisions with little progress.
🏛️Shortcut #5: Use existing decision-making frameworks
Many different types of frameworks and tips exist to help you make better decisions: Inversion, Pre-mortems, Weighted Decision Matrix
Whatever framework or approach you choose, continuously try to improve the process of decision making. As you become better at making decisions you will be able to know what you can change and what you cannot. And, you will be able to better recognize what types of decisions to make.
The process of decision-making is more important than the outcome.
🤔 “How to Decide” Group
If anyone wants to join us in reading Annie Duke’s new book How to Decide, let me know by sending me a DM on Twitter @hollyhliu and I will add you to a group! If you don’t have time, I highly recommend listening to Annie Duke in this podcast.
❤️Sharing is caring
❤️If you think anyone could benefit from this newsletter or post, please share! ❤️
🙏Gratitude to Jen Liao for inspiring this topic. Special thanks to Jen and Karen Hong for reading drafts of this article.