Seed the users whose peers will see the product first

Seeding networked power users means getting the product to a small group whose work many peers watch, then letting that visibility do the selling. In the one case here, a newsletter author says Granola's users mostly arrived through venture capitalists or people they told. He credits the 10% weekly growth he reports to Granola's bot-free design, not the seeding.

How it works

The idea is to pick users whose peers will see the product before anyone else's peers do. Then make sure those peers can try it without friction. Here is how the steps map to the Granola case.

1. Pick a group that sits at the center of your buyers' network. Granola is a notepad that preps a brief before external meetings and writes notes, action items and follow-ups afterward. Its early users were mostly VCs or people connected to them, which the author of this newsletter analysis of how Granola grows calls a beachhead strategy.
2. Let adoption cluster inside firms. The same write-up lists Lightspeed, Lux Capital, Benchmark, Sequoia, Accel, USV, Firstminute and Betaworks as firms where Granola caught on. The author says he started using it around summer 2024.
3. Decide how the product shows up in front of peers. Granola transcribes in the background, with no bot joining the call. That removes the in-meeting advertisement a bot would create, so spread depends on users talking about it and sending notes.
4. Give peers a free way in. Granola's notes stay private unless the user chooses to send them, and its homepage says free use is unlimited, with a 30-day window on older notes. You can see the positioning on Granola's homepage.

The write-up also reports Granola's claim that more than half of its users hold leadership roles, a figure the author says he has not verified. If that claim is accurate, it fits the pattern: senior people are visible to many others.

This is a loop, not a launch. Seeded users bring in peers, and peers in the same firms bring in more. The growth loops guide shows how to draw yours.

What it takes

Our estimate: about 40 hours of work, and about 30 days before you can read a signal. Most of the hours go into building the seed list, doing personal outreach and onboarding calls, and adding source tracking such as invite codes or a signup question. A smaller share goes into a share path that lets a seeded user send output to a peer. The signal takes a month because peers have to see the product in use a few times, then sign up, then stick for a couple of weeks before you can tell referred users from curious ones.

What to measure

The first two metrics show whether signups follow seeding, though only a holdout comparison can show cause. If the share of signups naming a seeded user does not rise above your baseline, the group is not as central as you thought. Signups per seeded user, compared with ordinary active users, tells you whether these people carry more reach. You can turn that into a k-factor with the calculator.

Retention of referred signups is the quality check. A good reading is referred users retaining at least as well as other new users. A bad reading is a spike of signups that fades, which suggests curiosity rather than fit. Read it by cohort, using the retention curve guide.

Accounts with several active users show whether adoption clusters inside firms, as it did for Granola. Seeded users' own eight-week activity matters because a seed who stops using the product stops showing it to anyone.

When it backfires

• You seed people who are loud but do not use it. Reach without daily use produces mentions, not signups. Check activity before you count anyone as a seed.
• You credit the wrong cause. Even the source for Granola's growth figure ties it to the bot-free design rather than the VC seeding. If the product is the real driver, the seeding plan will not travel to your next segment.
• The product is invisible in use. Granola's no-bot design means peers on a call do not see it working. If your product is similar, you need a share path or the loop stalls.
• The seed group is too narrow. A tool that becomes known as being for one profession can struggle to cross into adjacent buyers. Plan your second ring before the first one saturates.
• You rely on unverified claims. Granola's leadership-role figure is the company's claim, reported secondhand and not checked. Do not plan around numbers like that.

Run it as an experiment

Hypothesis: seeding a small set of central users in a target segment will raise peer signups and multi-user accounts in that segment, compared with similar segments we do not seed.

Comparison: we recommend a holdout by cluster. List 10 to 20 firms or communities of similar size, seed half, and leave the rest alone for eight weeks. If you cannot split, run a before-and-after with a fixed four-week window on each side and no other launches inside it.

What to log: signup source (invite code, referral link and a one-question survey), the firm or email domain, the seeded user who referred them, and week-1 and week-4 activity for every signup.

Decision rule: we suggest keeping the program if seeded clusters show clearly more signups per firm and referred users retain at least as well as other new users. Drop it if signups rise but retention falls. Size the clusters first with the sample size calculator, and write it up using the growth experiment guide.

How strong the evidence is

The grade is practitioner report. Every number comes from a third-party newsletter, not from Granola itself. The only company source is the homepage, which describes the product and plan, not results. The case outcome is mixed. The write-up does not show a baseline, signup sources, conversion or retention by channel. Its 10% weekly figure is credited to the product's design, not to seeding. The leadership-role claim is unverified. Treat this as a pattern worth testing, not a proven result.