Running the Company Week to Week

Most early startups do not fail from one bad decision. They drift. The weeks fill with work that feels productive, nobody watches the numbers, and months pass without a clear answer to whether anything is working. A simple, steady operating rhythm turns effort into progress, and it costs far less than the drift it prevents.

One Metric That Matters

Pick one north-star metric that best captures whether customers are getting value from your product. For a marketplace it might be completed transactions; for a business tool, weekly active teams; for a subscription product, paying customers retained.

The metric should move when customers succeed and stay flat when they do not. Avoid numbers that are easy to inflate, such as signups or page views, unless your business really depends on them. If the north star can rise while customers are unhappy, it is the wrong metric.

Then choose three to five input metrics that you believe drive it and that your team can influence directly, such as new trials started, activation in the first week or sales calls held. The north star tells you whether you are winning. The inputs tell you what to work on this week.

Put these where everyone can see them. Revisit the choice every few months, not every week, or you will never have a trend long enough to learn from.

If you are unsure what to pick, Paul Graham's advice in Startup = Growth (2012) is a good default: the best thing to measure growth in is revenue, and the next best, for startups not yet charging, is active users. Y Combinator measures growth per week, and Graham puts a good rate during the programme at 5 to 7% a week.

A Weekly Operating Cadence

A week suits an early company. It is short enough to correct course quickly and long enough to ship something real. A simple cadence:

• Monday, plan. Review last week's numbers, choose the two or three things that matter most this week, and write them down.
• Through the week, ship. Protect long blocks of time for building, selling and talking to users. Release small changes often rather than large ones rarely.
• Friday, review. Look at the north star and inputs, what shipped and what you learned, and note what you would do differently.
• Every week, talk to customers. Schedule conversations with users, prospects and people who churned. Founders often drop this habit first, and losing it is a common reason companies stall.

Keep the meetings short and the notes in one shared place. What you are building is a feedback loop: plan, act, measure and adjust, every seven days.

Monthly Updates and Fast Decisions

Once a month, send a written update to investors, advisers and anyone else with a stake in the company. Keep it brief and use the same structure every time:

• The north star and key inputs, with last month's figures for comparison.
• Cash in the bank, monthly burn and runway.
• What went well, what went badly, and what you are focusing on next.
• Specific asks, such as introductions, hires or advice on a decision.

Send it in bad months too. Investors who only ever hear good news stop believing it, and the ones you tell about problems early are often the ones who help. Writing the update also forces you to look at your own company plainly.

For day-to-day decisions, borrow a framing Jeff Bezos set out in his 2015 Amazon shareholder letter. Some decisions are one-way doors: irreversible or nearly so, and worth making "methodically, carefully, slowly." Most are two-way doors: if you get them wrong, you can walk back through. Bezos wrote that these "can and should be made quickly by high judgment individuals or small groups," and warned that growing organisations tend to apply the slow, heavyweight process to most decisions, two-way doors included, which leads to slowness and too little experimentation. Make two-way-door decisions quickly and treat them as cheap experiments. Save deliberation for the few choices that lock you in, such as co-founder equity, a large lease or a major architectural bet.

Persist or Pivot

Every founder eventually has to decide whether to keep going or change direction. No formula settles it, but some signals help.

Persisting is usually right when customers are actively using the product, some are paying or clearly would, and your inputs are improving, even slowly. Slow growth with strong retention usually points to a distribution problem, and the idea itself may be sound.

A pivot deserves serious thought when, after real effort, users try the product and do not come back, nobody will pay, and conversations keep turning up a different and more urgent problem. The strongest pivots come from something you learned from customers, rather than boredom or a new trend. Slack shows the same instinct applied to evidence from real use. Stewart Butterfield's company Tiny Speck built it as the software the team used to create and run its online game, Glitch, because off-the-shelf tools fell short for a team spread across several locations. When the game closed in late 2012 for lack of players, the tool the team relied on every day became the product.

Ignoring these signals is expensive. In a 2026 analysis of 431 venture-backed startups that shut down since 2023, CB Insights found poor product-market fit was cited in 43% of cases.

Set a review point in advance. Decide now what evidence you expect to see by a given date, and when the date arrives, compare honestly. That keeps the decision from being made on a bad day, or put off indefinitely on a good one.

Cash and Runway

When cash runs out, the company ends. In the same CB Insights study, "ran out of capital" topped the list at 70%, though the authors note it is almost always the final cause of death rather than the root problem. Know three numbers at all times: cash in the bank, monthly net burn and runway in months.

Then ask the question Paul Graham poses in Default Alive or Default Dead? (2015): assuming expenses stay constant and revenue keeps growing at its recent rate, will you reach profitability on the money you have left? Graham observed that half the founders he talked to did not know, and advised starting to ask too early rather than too late.

Review the numbers at least monthly, alongside a simple twelve-month forecast. Update the forecast when something changes, such as a hire, a large contract or a price change. Keep company money in a separate business account, and check recurring software charges, which pile up unnoticed.

Start fundraising, or cutting costs, well before you need to. Raising usually takes longer than founders expect, and negotiating with little runway left weakens your position. If the forecast shows trouble ahead, act while you still have options.

Focus and Saying No

Founder attention is the scarcest resource in a young company. Feature requests, partnerships, conferences, advisory calls and side projects all compete for it, and most of them sound reasonable on their own.

Measure each one against the north star and this week's priorities. If it does not clearly move them, the answer is no, or not yet. Keep a "later" list so good ideas are captured without being acted on.

Saying no to customers is hardest. A request from a paying user feels urgent, but building every request turns your product into a pile of one-off features. Listen for the problem behind the request, and build for the pattern you hear across many customers rather than for the loudest one.