Time to Market

Time to Market S01E09 – Making Tough Decisions as Founders

- - - - -

Sean K. Murphy: This is episode 9 of the Time to Market podcast. Today, we will talk about making tough calls. One of the keys to startup success is making good decisions, and we're going to look today at a couple of frameworks for that, and share some of our experiences of what we've seen work and not work.

Étienne Garbugli: A lot of founders in the past few years have had to make some really tough decisions with the downturns and the funding environment. So, it's been a bit of a tougher go for the last few years.
Sean K. Murphy: So, when you think about what constitutes a tough call, what are the elements or aspects of a decision that make it harder?
Étienne Garbugli: One idea Jeff Bezos has been pushing is reversible versus irreversible decisions. Some decisions will be very difficult to reverse as the company moves forward, especially if it's going fast.
It's important to realize that some decisions that structure the company moving forward cannot be changed. In a fast-growing startup, you may need to make them before you have a well-defined decision-making process.
Sean K. Murphy: I see two things that make startup decisions harder. The first is the need to proceed with limited information. You can get run over when you are small if you stand still. But you can definitely make poor choices. The second is that you may face unfamiliar challenges or proceed into essentially unknown terrain.
Étienne Garbugli: Yeah, it's always the new problems you didn't foresee or didn't know could exist that get to you. I have worked in many startups and made a lot of mistakes. I can recognize common patterns, but when you get "off the map" where your past experience is less helpful, it's challenging to get a clear understanding of what's going on and what might happen next.
Sean K. Murphy: I don't buy into Bezos' framework of irreversible versus reversible decisions: I view all decisions as the irrevocable commitment of resources, so all decisions have an irreversible aspect. But some decisions clearly commit more resources or entail more risk of negative consequences.
When you are in a small team, you can pull everyone into one room–or one Zoom–and walk around the situation and identify options together. You can agree to "try something" and run a small experiment before making a larger commitment. As the company grows or the effects extend beyond your firm, you must ensure the appropriate people are consulted. It's often the case that a mistake is not the result of who was in the room but who was left out.
Lisa Solomon has this three-conversation model where you frame the problem in the first conversation–if people are left out, and you realize it, this can take two conversations. You can bound the problem with the right people in the conversation and reach a working consensus on your objectives and constraints. Now you can have the second conversation, which generates alternatives. Then you let those percolate a little, and in a third conversation, you decide and plan how to move forward.
We've talked about introverts and extroverts: this approach gives introverts more time to settle their thinking and share it than if you jammed it into a 45-minute meeting. If it's a minor decision, you start and make fine-tuning adjustments as you go: ready-fire-steer.
What have you seen work, Étienne?
Étienne Garbugli: There's something to be said for accepting a high failure rate for minor decisions.
The quality of your decisions is a crucial test for your startup. As you scale your organization, you are building a machine to make decisions that include an evaluation cycle for the results you achieve.
I can see your point about decisions being irreversible, but if you have an ongoing evaluation process, you can adjust your plans and improve your decision processes.
One model that I think a lot about is the OODA loop from John Boyd: Observe, Orient, Decide, and Act. You gain most of your momentum in the orientation phase as you make sense of the information you have in the context of your experience.
I like to view decisions as experiments where you set time limits for results and reassess whether the plan worked. Those assessments are part of the evaluation cycle and trigger the correct behavior over time.
In some ways, the ability to recognize problems and readjust plans is more important than worrying about making a mistake when you first make the decision. Unfortunately, many organizations don't take the time to reassess past decisions and find it hard to improve how they make decisions. They build on past mistakes instead of making adjustments and lock the company into a poor trajectory.
Sean K. Murphy: I like your perspective that founders are building a decision-making machine. One mistake I see some founders make is that they need to make their decision rules clear.
It can be enjoyable for many decisions to come to the executive team. I had an old boss who ran a business unit at Cisco for a while. He told me that he had an epiphany one day. He realized that whenever an easy decision reached his desk, there had been a failure in the chain of command because he should only get the tough decisions.
Easy decisions should be made more quickly by lower-level managers or employees. If they reached him, he had not made his criteria or decision rules clear, or he was grabbing onto something because he was worn out from the tough calls and wanted to feel good that he had solved a problem. But that was not true leadership.
Étienne Garbugli: You have to be careful not to hire people who are all too willing to delay decisions. Some people would rather be told what to do; as a founder, you should take great care not to hire them.
A good test of the founders' ability to scale an organization is how effectively his team can operate day-to-day on its own. Can he delegate decision-making to the appropriate groups or departments and enable them not only to make good decisions but also to re-evaluate them as needed to make adjustments and better decisions. It's a balancing act; you must give people leeway to make mistakes and learn. If you do it well, you can build a lot of momentum.
If you look at how the Ukrainian army delegates decision-making to units based on accepted principles, you see a good model for startups. The Russians are very much top-down and becoming more paralyzed as the Ukrainians operate inside their decision cycle.
In general, a team with shared values and high trust will outperform one where all the decisions have to flow to a manager. But you need mechanisms that encourage delegation based on clear principles and guidelines. You need to reward the behaviors you want, promote the people who engage in them, and fire those who don't.
Sean K. Murphy: I think the approach of pushing a decision to those with first hand information works well if you also provide them with guidance and resources. There is a phase change that takes place when the whole team cannot fit around a table and have a conversation. It’s somewhere around 16 to 24 people. If you can start to delegate decisions before you hit that headcount you are more likely to stay nimble and responsive. You have to make your principles clear and tolerate mistakes made for the right reasons.
Étienne Garbugli: But in startup situations where there's a tough call: things are not exactly on plan, your runway is shrinking, and time is running short, how should founders decide? How do you advise founders to make the right decision, or at least the best one they can, given the information they have?
Sean K. Murphy: Many bootstrappers would call that Monday morning.The first factor to consider is how much time you have. If there is a fire in the building, you must get everyone out now. So that's a critical and urgent priority. Leaders sometimes let situations drift into crisis because they don't set a trigger even though they know it's critical.
The people who scare me are those who say, "Never let a crisis go to waste." They can become arsonists and saboteurs because a crisis provides a focus for action.
If it's not a fire in the building, ensure you have the right people in the room. This can involve reaching out to advisors, other entrepreneurs, or anyone with relevant experience who can offer insights. You need to ensure you are considering at least three to six viable alternatives.
Sometimes a decision gets framed as "pivot or persevere" or a binary choice. If you are only considering two options and it's a significant challenge–or opportunity–you haven't done enough to generate options and will likely make a mistake.
Two tests: review your process to ensure that you have broad input and that other folks with relevant knowledge have helped you look at several alternatives. And even when the situation is serious, allow for a certain amount of deliberation.
Étienne Garbugli: But that approach requires that people are willing to contribute. If there's going to be a free sharing of different alternatives people need to be willing to challenge solutions that leaders have proposed. They may need to go counter what some on the team may believe about the sources and extent of the crisis, and to disagree with what they are advocating.
How do you get people to speak frankly in a crisis? How do you get them to suggest more options when everyone feels the need for quick action? I often see a pattern where people either want to “stay the course” or jump on the first alternative that’s proposed. A crisis already makes people tense, few may feel involved enough in the decision-making process to want to “rock the boat.”
So how do you make sure you are getting this wide range of perspectives or rich set of options that you say are essential to choosing a good strategy?
Sean K. Murphy: Good point, “it’s very easy to say, very hard to do.”
You have to make it clear that you need their help and actively encourage feedback, suggestions, and a free exchange of ideas. You can't open with, “Here's what I think we should do but I am but I'm interested in what you think.” Because that's what you say when you want to let people know that you have made the decision and are announcing it gently: “This is what's coming so get ready, right?” Which is a kind of fake democracy.
I managed the engineering services group at Cisco, which was about 15 people at the time. The company was growing rapidly; I had joined about two and half years earlier and had hired almost everyone in the room. Everybody liked the stock splits and rising price, but rapid growth puts a lot of strain on an internal services team.
So people were unhappy but wanted to make things work. I have a strong personality, so I put one of my supervisors in charge, who was well respected and left the room. I wanted them to take an hour, agree on a prioritized list of issues or problems, and write them on a large whiteboard so that we would discuss them together. True story.
About 45 minutes later, someone came and got me. There were nine challenges listed on the board in very frank language. I thanked everyone for their candor and told them I appreciated their commitment to improving things, which I assured them I was also committed to.
So we started to walk through the items, and in each case, I asked questions to clarify what they meant and explore particular examples.
The first bullet was, "We take too much shit." As we peeled the onion on this, it meant that I was not standing up enough for the group in negotiations with other departments we served. They cited examples where I had made mistakes. I changed my approach to certain situations; we changed several of our internal processes and procedures.
If you ask for feedback, you must understand the meaning of someone's remarks or suggestions. Ask questions like, "Tell me more about this," or "Can you give me an example." Avoid the temptation to deflect or object. Instead, acknowledge what people are telling you. You don't have to agree, but you have to demonstrate that you understand.
I know this is at odds with the common portrayal of the visionary founder or dynamic CEO, but it's what you need to do to harness a team's collective problem-solving creative talents.
Étienne Garbugli: After your team generated these different options, how did you decide which to implement and make them work in the organization?
Sean K. Murphy: Final decisions took a period of weeks, there wasn’t a single dominant problem. It wasn’t like the Gordian Knot where you get your sword and untie everything with a single well-placed cut.
I remember one fundamental mistake clearly because it was my fault–I had mistakenly overconstrained how we debugged new component models. We relied on design prototypes–first iterations of new products–to debug our library components. Software startups can make a similar mistake when they rely on beta sites or early users to find most of their bugs.You inject problems into your customer’s workflow.
I decided we were going to fund test boards out of our own budget to verify certain components that were complicated and hard to get right. Also, in some cases we had followed the datasheet specification for a new part and it had been wrong. At the time it was only a few thousand dollars and it kept the vast majority of our mistakes off of the critical path for a new product. We could do our test vehicle weeks in advance and even do a second spin if problems were detected.
For whatever reason I had had a mental block against doing this. So if you are in charge, make sure that you are not adding artificial constraints to a problem. If you ask the group to help you map out options you can often detect where you have artificially constrained the solution space, or where you should be looking for opportunities.
Étienne Garbugli: Too often, we focus on solutions before clearly defining the problem. Founders tend to take more ownership of problems, which is usually a good thing, but it can inadvertently deny employees agency or ownership. Smart founders want employees to bring their brains and passion to work, not just do what they are told.
How did you decide the priorities for the nine items?
Sean K. Murphy: A mix of the difficulty level, need for coordination or negotiation with other groups, cost, and impact on the quality of results we delivered or contributed to. The nine were more or less decoupled from each other, so we started with the ones that only needed the team to act. Others that required external coordination had to begin with discussions with other groups who would be affected or whose help we needed.
You mentioned ownership, and it reminded me of a talk I heard by Sanjay Srivastava after he had bootstrapped Denali and sold it to Cadence for several hundred million dollars. One thing he said stuck with me, "Whenever I heard at Denali that we had a sales problem, I immediately challenged the engineers to see what they could do to address it."
There is a natural split between engineering and sales. The engineers--I say this as a recovering engineer--tell themselves, "We have built this beautiful, powerful thing; it's a shame the sales guys can't sell it."
Sanjay helped the engineering team appreciate what he had learned in the early days of bootstrapping Denali. When he challenged them to take ownership of the problem, the engineers often discovered that they could modify the product in ways that made it much easier to sell.
We draw these lines between groups and say, "The problem is on your side of the line." If you can get folks together to look at a challenge holistically--do a few orbits around it and see it from different sides--you will often uncover new options.
Alas, brainstorming is necessary but not sufficient. At the end of the meeting, it's not enough to announce, "These are some great ideas guys, go make it happen," leave the room and expect something to happen.
All those wonderful ideas need to get organized into project plans, deadlines, commitment-- and, to your earlier point, evaluation checkpoints to assess interim and final results.
Étienne Garbugli: How did you tell that your changes were working? How long did you wait and how did you make sure that you were getting the results you were expecting?
Sean K. Murphy: In the case of the separate test build for the components, there were two evaluation criteria. First, did the test board catch problems? If it did, we knew we were at least catching some issues early. Second, were we still seeing component errors impacting the prototype system bring-up? If we were, we needed to upgrade our test board approach to make it more potent.
You need to map your dependencies and possible decision points. You need to be able to compare "before and after" to assess the potency or impact of a new approach. One mistake I see many entrepreneurs make when they try to obsolete an established workflow is that they don't plan for the customer to return to the established path if the new approach is unsatisfactory.
We ran the test board example early enough to identify problems without impacting the prototype build. We did not introduce a new dependency on the test board working.
Étienne Garbugli: So I hear two layers in the decision process: the specific tough call you need to make and an assessment of how the organization can implement it or may react to it.
The top layer is your overarching decision-making model, which determines when things are difficult and need to be discussed in a group with an agreed process. The bottom layer is the situation's specific characteristics that need to be evaluated and managed. I recognize that the top layer is intertwined with the need to improve the way you make decisions and re-evaluate the impact of specific decisions as they play out. The bottom layer benefits from improvements in your ability to make sense of the world and diagnose underlying forces at work.
We’ve covered a lot, what are your specific takeaways?
Sean K. Murphy: You make some good points. I have two take-aways.
Maintaining team trust and morale is critical. When there's a lack of goodwill, perhaps because people affected by the decision aren't consulted, you are less likely to get the results you want as things roll out.
You need to map out who needs to be consulted and who needs to be informed. When your team can fit around one table, people are more inclined to ask questions and offer suggestions. But as the organization grows, it gets easier to leave people out. The founders are still in the thick of things, so it can be hard for them to appreciate the need for communication, coordination, and collaboration with the right team members. They know what's happening but don't realize that others who need to don't.
Étienne Garbugli: I would add that you need to communicate the decision and your reasoning. If you explain the key alternatives considered, your criteria, and your decision process, you help foster a shared decision-making framework that enables more effective delegation. You help team members improve their mental models and broaden their perspective so that they consider more than their needs and look at how others may be affected.
Explaining your principles helps junior staff and people with less experience making tough calls improve. It cuts down on politics, which can otherwise slow decision-making and hinder effective action.
Sean K. Murphy: I like your takeaways because they recognize that you face two challenges: survival and the need to improve decision-making capability at all levels of the organization. A "decision record" captures options considered, decision criteria, what you expect will happen, and at least one checkpoint where you revisit what results were achieved. As you said earlier, it's less about the decision and more about the process you used and ensuring you follow up.
Étienne Garbugli: With each tough call you make helps build your decision process and operating model. However, people watch what you do more than you say. If you are dictatorial in a situation so that your stated values and actions conflict, people will follow your actions.
Sean K. Murphy: The truly tough calls are dilemmas that put two values or critical objects in conflict. We want low cost and high quality. We want a rich feature set but fast time to market. These are dilemmas with no right answers.
Étienne Garbugli: That's a good, good place to end this discussion. So, thanks for listening. You can reach out to us and give us feedback @leanb2b on Twitter or @skmurphy on Twitter, and we will see you guys in the next episode.