My friend, Vijaye Raji, founded a startup, Statsig. The four values he set for the company have one thing in common: they are all focused on execution.
Statsig, much like its values, aims to help companies move fast and make data driven decisions. They make all the tools we used and loved inside Facebook accessible to everyone.
This may strike you as curious—shouldn’t smooth execution be a given for any company? But you would be surprised at how easy it is to get bogged down by inefficiency.
I often think back to the early days of PayPal, when it was exhilarating to come into work every day and see the changes we made ship quickly. The momentum we built up was inspiring, and getting to constantly see our progress made our efforts feel impactful and worthwhile.
Then something changed. Things that used to take hours started taking days. Things that once took days started taking weeks. Projects were becoming multi-quarter affairs with planning and involvement from many different teams. At one point, the scopes got so large that one engineer joked with me, “It takes three days to go to the bathroom around here.” I was so frustrated, I ended up taking all of the projects we had recently scoped and pulling the original estimates. The inflation of the cost and the timelines compared to what we had initially planned for was staggering.
I wasn’t the only one who was frustrated. One day, Hugo Olliphant, another early PM, stopped by my desk to vent. He had discovered a content error on the site, which literally said the opposite of what we meant. He spent days hunting down people to address it, hitting wall after wall. There were issues with translation (string extraction, I’m looking at you), then with QA, then with an escalation to determine if it was worth a hotfix on the live site. It took three weeks for this minor content error, which should have been a 15-minute change, to be addressed.
Many issues can slow down execution in a company. Often, these issues are small and seemingly innocuous, but over time, they build up to the point where they prevent things from getting done.
I have worked at many incredible tech companies that did a lot of things right, but this post is about the things that can go wrong. Execution problems can happen even in very successful companies. These lessons remind us that no team and no company is infallible. It is important to always keep an eye out for things that degrade execution and take your focus off the customer you are serving. Here are a few examples:
1. Too many cooks in the kitchen.
I once worked on a special research project for the executive team to try to figure out why execution was so slow. I spoke to a dozen teams, and it turned out that there were simply way too many parties involved in every decision.
One team said, “We started with a dozen core team members, and now the group has over 250 people weighing in.”
Another explained, “Everyone has an opinion, and no one agrees with each other. It is not clear who gets to have a say in what, and ultimately who makes the call.”
To avoid this, it’s important to ask, “What is the minimum-viable team necessary to make this decision?” In the DACI model, only the Driver, Approver, and Contributor typically need to be in the room for a decision. The Informed should be kept in the loop, but not involved in the actual decision-making process. There is a temptation to be overly-inclusive and have everyone involved, but things become unwieldy once you have more than a couple dozen people engaged, all with different opinions.
2. Secret vetoes.
If I asked you who could say no to something at your company, what would be the response? The answer the leaders give and the answer a product team gives may be completely different. This is the trap of secret vetoes, which are a major threat to execution.
I once had a team member tell me, “We say we disagree and commit, but actually we are agreeing not to commit.” Multiple people were secretly filibustering the project but not saying it out loud. They did everything in their power to slow things down, in small and big ways. These secret vetoes made the project drag on for months and months longer than necessary.
In some companies, there are groups whose roles are to manage risk, such as Legal. Many of the meetings with Legal teams are sessions warning of the dire risks the company faces if it proceeds with an action.
By working with the SVP of Legal and General Counsel, Greg Packer, we were able to turn this model on its head. After they share the risks and mitigations, I ask each lawyer present, “Given the risks you outlined and the opportunity, as equal leaders at the company, do you think should we proceed?” Nearly every time, the answer has been yes. The difference is that they co-own the decision and risks with the rest of the team, rather than being on the outside.
3. Risk aversion.
Some companies have a “must-win” attitude. Teams are not allowed to lose or fail, and those who do are punished.
Imagine a company that says, “We want every product we ship to be successful.” That means building in contingency plans, redundancy, and failsafes. It also means months of planning, tons of testing, and perfect execution. This often results in a team that can ship only one thing each half. The company would be much better off if they tested four things and had a 50 percent hit rate, thus doubling their absolute success. Or what if they could test ten things with a 40 percent hit rate? In this scenario, there are more teams that “fail,” but the company and its customers are substantially better off.
A recent massive growth win at Ancestry was only possible when the first iteration failed and the team retested it by adjusting the user experience. Failure actually led to success, because the learning machine was getting more data through and iterating faster. By decriminalizing failure and building a system of learning, a company can not only succeed at a higher rate, but it can also learn from the things that didn’t work.
4. Poor goal-setting and accountability.
I once led a team where the numbers were looking a bit anemic, so I asked the team to figure out why we were behind on our goal. In order to hit the goal, they rolled it out further, thereby increasing absolute engagement by widening the funnel, but without understanding why retention was weak.
If you reward absolute numbers, there are many ways to achieve them… poorly. Pouring more users into the top of the funnel without fixing engagement and retention can help you hit a short-term goal, but at the cost of long-term success.
Remember, you get more of what you measure. Being mindful of what you want more of and what you want less of, and understanding how the numbers can be gamed, is part of the culture of ownership you want to create.
5. Slow escalation.
At many companies, escalations are taboo. Teams try to work things out amongst themselves, leading to compromised products and delayed timelines. The problem is that the lack of alignment actually exists beyond the control of the teams. No amount of individuals working together can fix a fundamental disagreement between two leaders or functions without a healthy escalation process.
One day at Facebook, one of my PMs reached out to me. She had been negotiating with another PM for weeks, trying to ensure that a change on their ranking would not negatively affect our product. At her wits’ end, she finally came to me to ask for help. I reached out to the head of the other team. Within a single discussion, we had found a resolution that didn’t delay their launch and also protected our product from losing distribution.
I admire LinkedIn’s five-day escalation technique as a way to force resolution quickly (video version). Long and drawn out fights reduce productivity, strain relationships, and ultimately result in compromised products.
6. The gauntlet is exhausting.
I once spoke to a PM who was working on a key company priority. His words echoed those of Hugo: “We have meeting after meeting, and people don’t seem to be on board. It is exhausting rather than energizing working on this project.” Imagine you are a PM leading your team through this kind of a gauntlet. Each time you clear a hurdle, a new one mysteriously arises in front of you. Not only does it slow you down, but it’s also demoralizing for you and your team.
Many years ago, when I was at PayPal, I helped draft our spec template. At the back of the template were a couple dozen teams (e.g. Legal, Risk, Customer Service) that you had to get clearance from before your spec was approved for coding. If Team 14 objected and requested a change, you had to submit a CR, and the first 13 teams had to reapprove it. This meant that getting anything shipped was an endless cycle of requests and reapprovals from everyone on the list. The process took months and placed a burden on anyone trying to make progress. The more you can streamline these processes, the more you can maintain momentum and keep moving forward.
7. Constant changing of priorities.
Teams need time and space to execute. We can come up with new ideas faster than it takes to code them. Constantly changing your area of focus wastes cycles and ends up requiring a ton of rework. The worst part is that often these changes are not clearly documented, and thus, teams are whipsawed as they try to execute.
One of the things we decided to do at Ancestry was to create a company-level product roadmap starting in H2 2021. Before, each team had a different roadmap, and it was hard for anyone to get visibility into the big picture of what the company was working on. Changes happened within teams, so those who had to support the launches lacked clarity about when changes happened or projects moved. For 2022, we decided to create a two-year roadmap with owners for the four key workstreams. This meant that our priorities would be clearly published and shared widely, and everyone had access to them. It also meant that changes or delays would now be documented and justified, rather than slipped, to prevent ripple effects on other teams.
8. Lack of ownership.
One of my top priorities when I first arrived at Ancestry was to understand how to get things done within the company. During my listening tour, many people mentioned a feature that was in high demand but getting blocked. I was forwarded old specs from 2018 talking about what it would take to roll it out. I spent a few weeks trying to understand what the blocker was, then I set up a meeting with the GM of the business, our CTO, and our VP of Legal. In 20 minutes, we unblocked the project, and it shipped starting during the holidays.
I didn’t have any silver bullet or any particular insight that helped the project get unstuck. The only thing I did was convene the small group of people who had to agree on how to ship it and get them in alignment. What I later realized was that even though many people had been working on parts of the project, no one person clearly owned it. The diffusion of responsibility meant that dozens of people played important parts, but there was no conductor for the orchestra.
Ownership means accountability. One team I used to work with liked to ask, “Who is the throat to choke?” I found that imagery uncomfortable, but it was memorably illustrative. Every project has to have someone who is personally on the line to see something through. If everyone is leading it, no one is owning it.
9. Rewarding the wrong things.
One of Statsig’s values is Results Over Effort. The reasoning is simple: rewarding motion over progress is one of the most insidious traps against execution. If you are rewarding effort, you may have a lot of motion, but not necessarily a lot of results.
Likewise, there are also companies that value relationships over impact. It’s true that relationships are critical to company success. The problems arise when people are not willing or able to make difficult decisions necessary to serve the product or customer, whether because they will hurt another team or upset someone. This becomes a hindrance, because it disincentivizes people from taking action and being decisive.
10. Too much incrementalism.
When executing, some teams are encouraged to make small, iterative changes rather than big ones. My philosophy has always been the opposite: start big, testing to figure out the maximum change you can ship. If it doesn’t work, then you have your answer. If it works, but you find it is too aggressive, you can ratchet back accordingly, without trying to find the right balance point by making tons of slow, incremental changes.
When testing something, it’s critical to go all the way to the furthest you would go, where you can see the maximum change you can make at any one time. This allows you to understand the limits of your product. When teams are encouraged to test something piecemeal, the process ends up being drawn out, and may even result in wasted time.
What if you launched the most aggressive test you could and found out the change did nothing? On the flipside, what if you found out it increased conversion by ten percent? In either case, wouldn’t you rather know the result now, rather than after six months of tiny changes? By making the maximum change possible, you can gather much more information in a shorter amount of time and get a clearer idea of whether or not to move forward.
One final note: Statsig’s “No Sacred Cats” is in many ways my favorite of their values. While it’s not something that directly gets in the way of execution, I still think it’s worth mentioning here. When Vijaye and I were working together, our team’s mascot was a cat. We were a team that tested a lot, and many times things that tested poorly in one half turned into wins just a year later. The hashtag #nosacredcats was a reminder that we should always be learning and willing to challenge ourselves. We were never allowed to say, “We tried it before, and it didn’t work.” You would be surprised how many wins came from previously failed tests.
When teams complain of struggles with execution, we too often look to the individuals and ask why they aren’t doing better. In reality, execution is more often about culture, processes, and values than it is about individuals. As you build your companies, products, and teams, keep in mind the many things that can get in the way of execution. Unblocking your teams and unlocking your full potential means identifying and addressing these barriers.
Thanks to my friend, Vijaye, and his company for letting me use their values for this post. Statsig, much like its values, helps companies move fast and make data-driven decisions. They made all the tools we used and loved when we were builders inside of Facebook accessible to everyone, and if you want to learn more, click here.
How small things slow down the achieving of big thingsRead MorePerspectives
