Mistake #8: I Planned for the Work. Not for the Reality Around the Work.
This is the eighth post in a series about the mistakes I made starting up, and what I’d tell myself if I could go back.
Every founder underestimates timelines. This is so common it barely registers as a mistake anymore. Of course things take longer than expected. Everyone knows that.
Knowing it and actually accounting for it are different things.
I knew it. I still built plans that assumed the work would take the time the work takes, and nothing else. And every time, the something else was what undid the timeline.
What the Plan Always Assumed
When you sit down to plan a quarter, or a sprint, or a product launch, the plan is a model of the work. Feature A takes three weeks. Fundraising closes in six. Hiring the engineer takes a month. Customer onboarding is two weeks per account.
The model is internally consistent. The numbers are not invented. They reflect real estimates from people who know the work.
What the model does not contain is any of the following:
The co-founder conversation that needs to happen and keeps getting scheduled and moved. The investor who seemed close and then went quiet for three weeks and needed four more touchpoints before saying no. The key engineer who got sick during the most critical sprint of the quarter. The customer who was ready to sign and then their internal champion changed roles. The compliance filing you forgot was annual. The integration that broke in production the week of the launch. The advisor whose input you needed who was traveling for two weeks.
None of these are unusual. None of them are bad luck in any meaningful sense. They are the normal texture of running a company. They happen every quarter, to every team, in some combination.
And yet the plan never made room for them.
Why We Do This
Part of it is optimism, which is not entirely a flaw. You need some conviction that the thing will happen in order to commit to it. A plan built on genuine pessimism is not a plan, it is a hedge.
But part of it is something more specific. When you plan, you are thinking about the work. The work is concrete. You can estimate it, break it into parts, assign it, track it. The reality around the work is diffuse and unpredictable. You cannot put it in a spreadsheet. And so it does not go in the spreadsheet.
There is also a version of this that is particular to founders who came from large companies, which I did. At Microsoft, there were systems for almost everything. Legal had a process. Finance had a process. HR had a process. When you needed something from another team, there was a way to request it. Things moved slowly sometimes, but the machinery existed and you could rely on it.
In a startup, you are the machinery. And the machinery breaks in ways that the machinery at a large company does not. Your engineer is also your IT support. Your co-founder is also your lawyer for the day when you cannot afford the real one. You are the HR department, the finance team, and the product manager all in the same week.
Every hour you spend on those things is an hour that was in nobody’s plan.
The Specific Ways Timelines Break
After eight years of this, I have noticed that timelines do not break randomly. They break in predictable categories.
People dependencies. Almost everything in a startup depends on another human making a decision or taking an action. Investors, customers, hires, partners, regulators. You can control your own pace. You cannot control theirs. Every external dependency is a potential delay that your plan did not price in.
Context switching. A founder’s day is not a series of focused blocks. It is a series of interruptions that occasionally has a focused block inside it. The sprint that was supposed to take two weeks gets stretched because you spent four days that week on a fundraising conversation, a team issue, and a customer escalation that all needed you specifically.
The decision that was not ready to be made. Some things cannot move forward until a decision is made, and sometimes the decision is not ready. Not because of laziness or poor planning, but because the information needed to make it well has not arrived yet. Waiting for that information is invisible on the plan but very visible in the timeline.
Recovery time. When something goes wrong, and something always goes wrong, the recovery takes time that was not planned. The launch that broke needs a fix. The hire that did not work out needs to be undone and restarted. The pivot requires the team to context-switch from what they were building. Recovery is a real cost that almost no plan includes.
The emotional overhead of uncertainty. This one is hardest to name but very real. Running a startup involves a level of sustained uncertainty that takes energy to manage. There are weeks where the weight of not knowing whether it will work is genuinely heavy. That weight slows things down in ways that do not show up on any project management tool, but are completely real.
The Rule I Eventually Developed
At some point I stopped trying to plan accurately and started planning honestly.
The difference: accurate planning tries to estimate each task correctly. Honest planning accepts that the estimate will be wrong and builds the buffer in from the start.
The rule I use now is simple. Take the timeline that feels right after careful estimation. Double it. That is the number I communicate externally. Internally, the team still works toward the original timeline, because having a real target matters. But the number I commit to with investors, customers, or partners is the doubled one.
This is not sandbagging. It is not being conservative for the sake of looking good when you deliver early. It is accepting that the gap between planned time and actual time is not exceptional. It is structural. And once you accept that it is structural, the right response is to build it into the plan rather than be surprised by it every single time.
The corollary: when someone on the team gives you a timeline, add fifty percent before you write it down. Not because they are wrong. Because they are doing what you did for years, estimating the work and forgetting the reality around it.
What Honest Planning Actually Changes
When you stop underestimating, a few things shift.
You stop having to explain why things are late. Not because you are hiding anything, but because the timeline you committed to was realistic and you hit it. This sounds small. It is not. Repeated lateness, even with good reasons, erodes trust with customers, investors, and your own team. Realistic timelines, consistently met, do the opposite.
You stop making decisions under artificial urgency. When the plan is too tight, everything feels like a fire. You start cutting corners, skipping steps, making calls you would not make if you had more time. Honest planning gives you the time to make better decisions.
You also stop blaming the team for things that were never possible in the first place. A plan that assumes nothing will go wrong will always generate underperformance. A plan that assumes some things will go wrong, and accounts for it, gives your team a real chance to succeed against it.
The Deeper Issue
Underneath the timeline problem is a belief that is worth examining.
Most founders who underestimate timelines are operating from a mental model where the company is behind. There is a gap between where the company is and where it needs to be. And every day that passes without closing that gap feels like failure.
That feeling is what drives the aggressive timeline. If we just execute well for the next quarter, we can close the gap.
The gap does not close that way. Because the aggressive timeline generates the same miss it always generates, and the gap feels as large at the end as it did at the start.
The mental model that actually works is not “we are behind and need to catch up.” It is “this takes as long as it takes, and our job is to keep moving consistently.” That is not resignation. It is accuracy. And accuracy, it turns out, is what compound progress is built on.
Not speed that pretends the reality around the work does not exist.
The bottom line: Founders consistently underestimate timelines not because they are bad at estimating work, but because they plan for the work and ignore everything around it. The fix is not better estimation. It is honest planning that treats delays, dependencies, and disruptions as structural rather than exceptional. Double your timelines. Stop being surprised by the predictable.
What has consistently taken longer than you expected? I’d genuinely like to hear.
I’m Gopi Krishna, founder of Hyperleap AI, where we build enterprise-grade conversational AI but for small and medium-sized businesses across the world. This is part of an ongoing series on the mistakes I made in my first years as a founder, written for anyone thinking about starting something of their own. Subscribe to Second Order AI on Substack to follow along.