90-Day Growth Playbook

In one quarter you can find where your funnel leaks, test a few fixes properly and turn the ones that work into routines your team repeats. It takes a scored audit, a monthly dashboard, experiments that each have a win line and a stop date, and a written procedure for every winner.

I wrote this for founders and small marketing teams, from seed to Series B, who have paying customers and a few months of funnel numbers but no repeatable way to grow. It assumes you already know who you sell to and which one or two channels bring in leads. If you don’t, start with the Foundations Playbook, which covers positioning and choosing channels, and come back next quarter.

The quarter runs 13 weeks: a set-up week, called week 0, then 12 weeks to design experiments, run them and scale what won (Figure 1).

Week 0, days 1 and 2. Before you change anything, score the funnel you have. The Funnel Health Checklist has 16 items in four sections: traffic and awareness, lead capture, nurture and follow-up, and conversion to revenue. Rate each item 0 (no or very weak), 1 (partial) or 2 (strong), so each section scores 0 to 8 and the total 0 to 32 (Figure 2). Do it in one sitting with whoever runs sales; it takes about an hour.

The total tells you what kind of quarter to plan. From 0 to 10, the foundations are missing: fix the basics before you test anything. From 11 to 18, the funnel works but leaks: pick the two or three biggest leaks. From 19 to 24 it is solid, and the quarter is about scaling; from 25 to 32 it is strong, and the work is leverage, such as packaging what already works.

One item is stricter than you may expect: answering demo requests within the hour. The Lead Management Playbook explains why speed matters that much.

The checklist shows which stage is weak; it doesn’t show why. So also talk to five recent buyers and five prospects you lost, and ask what nearly stopped them. The Win-Loss Analysis Playbook has the questions.

Week 0, days 1 and 2. The dashboard is a sheet with one row per month, and the first rows are the last three months, so every experiment has a baseline to beat. It tracks three groups of numbers:

• Traffic and leads: sessions, leads (form fills, sign-ups and inbound requests) and the visitor-to-lead rate.
• Opportunities and customers: opportunities (qualified calls, demos or proposals), new customers, and the rates from lead to opportunity and opportunity to customer.
• Revenue and efficiency: new monthly recurring revenue (MRR), marketing and sales spend, customer acquisition cost (CAC) and CAC payback.

Figure 3 shows all 16 columns, from A to P, with Example Co.’s numbers and the formula for each calculated column. Two details keep the numbers honest. First, the formulas use plain division, such as C2/B2, rather than DIVIDE, which works in Google Sheets but is not an Excel function^[1]. Second, enter new revenue as monthly recurring revenue, not total contract value, and divide CAC by the monthly gross profit a customer brings in, not the revenue. David Skok’s SaaS metrics guide, the usual reference for months to recover CAC, adjusts it for gross margin and suggests recovering CAC within 12 months, while noting that businesses selling to large companies often take longer^[2].

Read the three-month totals, not a single month. With a handful of new customers a month, one deal more or less moves CAC by a quarter or more.

Week 0, days 3 to 7. Choose one primary goal for the quarter, written as a number and a week, such as opportunities a month by week 12. Then find the stage that holds it back: the conversion rate that is lowest against your own past, or the section the checklist scored lowest. I don’t use published conversion benchmarks for this; they mix industries, traffic sources and definitions, and your own trend is a better yardstick. Figure 4 shows which of the six experiments in Step 4 acts on each stage. Pick two that act on your weak stage.

Before you test anything, fix what is simply broken: a form that doesn’t reach your CRM, a dead link, a button that says nothing about what happens next. If you already know the answer, it is a fix, not an experiment, and it should be done by the end of the week.

Weeks 1 and 2. Six experiments cover the whole funnel. Each starts from a hypothesis you can prove wrong:

1. Landing page conversion lift. A clearer value statement and less friction will lift visitor to lead. Build a second version of your main page, with a new headline, subhead, call to action and proof, and split traffic between the two.
2. High-intent lead magnet and email. A focused resource and a four- or five-email sequence will bring more qualified leads and calls. Build the resource, a landing page and a short form; the emails deliver it, give a quick win, teach something deeper, show a customer case and make an offer.
3. Personalized mini-audit outreach. Short, tailored audits will lift reply and meeting rates. Pick 25 to 100 target accounts, spend five to ten minutes on each, and send two or three findings with a request to meet.
4. LinkedIn thought leadership sprint. Practical posts every weekday for four weeks will bring more profile views and inbound leads. Mix how-to posts, stories and teardowns, and reply to comments every day.
5. Pricing or offer reframe. A simpler flagship offer will lift the win rate and deal size. Write the offer as an outcome, a scope, a price and something that reduces the buyer’s risk, and update the pitch to match.
6. Referral or expansion prompt. Asking at the right moment will bring introductions and upsells. List happy customers, script the ask and send it after a good outcome.

Before anyone builds anything, write a one-page brief for each experiment: the hypothesis, one primary metric, a guardrail metric that must not get worse, the baseline from Step 2, the win line, the sample, the launch and stop dates, and one owner (Figure 5). The win line and the stop date matter most: without them, every result looks like a win.

For a test on your website, the sample comes from a sample-size calculation, and traffic decides what you can test. The smaller the lift you want to detect, the more visitors you need: halve the lift and you need roughly four times as many. On a small site, test bold changes on the busiest pages (Figure 6). Outreach, offer and referral experiments rarely reach numbers big enough for statistics, so compare the result with your baseline and the win line you set in advance, and treat a pass as a strong signal rather than proof.

Weeks 3 to 8. Launch the first two experiments in week 3 and the next two in week 5. Two at a time is about what a founder and one marketer can build, watch and read. If two tests touch the same page or the same people, run them one after the other, or you won’t know which one caused what.

Hold a 30-minute review every Monday: update the dashboard, move each experiment along the log and check that nothing is broken. Don’t read a result before its stop date. Checking a test every day and stopping the first time it looks significant can push the real false-positive rate far above the 5% the test was designed for^[3]. On the stop date, decide: scale it, kill it, or change one thing and run it again.

Expect most experiments to lose. At Microsoft, about a third of tested ideas proved effective, a third were neutral and a third were negative; at Google and Bing, only about 10% to 20% of experiments produced positive results^[4]. A killed experiment is a result, so log what it taught you.

Keep the log in one shared sheet next to the dashboard, one row per experiment: name, owner, start and stop dates, hypothesis, primary metric, status, result and what you learned.

Weeks 9 to 12. In the last four weeks, read the remaining experiments, then choose between launching the last ones and scaling what won. With a small team, scaling a clear winner usually earns more than starting two new tests.

A winner is only repeatable when someone other than the person who ran it can run it from a written procedure. For each one, write a standard operating procedure (SOP) on one page: the trigger, the owner, the steps, the templates, the time it takes and the number that shows it is still working (Figure 8). Then decide who runs it from now on: the team, a freelancer or agency, or software. Some winners can become a packaged offer instead, a fixed scope at a fixed price.

Put the SOPs, the dashboard, the experiment log and the list of untested ideas in one place: that is version 1 of your growth playbook. Plan the next quarter from it, starting with whichever stage is now the weakest.

What changes by stage?

What mistakes should I avoid?

• Testing what is simply broken. Fix it in week 0 (Step 3).
• Launching without a win line or a stop date. Write the brief and work out the sample first (Step 4).
• Reading results every day. Read them on the stop date (Step 5).
• Running all six experiments at once. Two at a time is what a small team can build and read (Step 5).
• Judging the quarter on one month. A few deals swing CAC; read the three-month totals (Step 2).
• Keeping a winner in one person’s head. Write the SOP before you scale it (Step 6).