Choose your first channel by testing, not guessing. Shortlist three channels where your customers already look for a solution, run each for one to two weeks with a small budget and one metric (cost per activated user), then put all your effort into the winner for 90 days.

This is for founders with a working app or a test page and no steady flow of users yet. The running example is JobBoard, an imagined job tracker for small plumbing firms.

In short:

  • Early growth usually comes from going straight to users, online and offline.
  • Search channels catch people already looking. Social channels interrupt people.
  • Measure activated users, not clicks or sign-ups.
  • Doing one channel well beats doing five badly.

Start where customers already look

Ask how your customer found the last tool they adopted. Plumbing firm owners might say: "a mate recommended it", "I searched on Google", or "a supplier told me". Those answers are your shortlist.

Lenny Rachitsky studied how the biggest consumer apps got their first 1,000 users and found that seven strategies covered all of them. The most popular involved going to users directly: online, offline and through friends. The same holds for small business apps.

The book Traction, by Gabriel Weinberg and Justin Mares, describes 19 channels and a "Bullseye" method: brainstorm all of them, rank a few, test three, and focus on one.

Compare channels by speed and intent

Direct outreach and search ads give the fastest signal. SEO is slow but compounds.

Channel Best when Time to signal Cost
Direct outreach Customers are businesses you can list Days Your time
Communities Customers ask for help online Days Your time
Search ads People search for the problem Days Ad budget
Social ads Customer is easy to target, not searching 1 to 2 weeks Ad budget
SEO and content Many people search, little competition 3 to 6 months Your time
App store search People browse the store category Weeks Your time
Partners Someone already serves your customer Weeks Revenue share

For JobBoard, direct outreach (calling firms from a local directory), search ads ("job tracking app for plumbers") and a supplier partnership make a sensible shortlist.

Search channels only work if people search. Look up your phrases in Google Ads Keyword Planner before you choose search ads or SEO. If the phrases get a few dozen searches a month, search won't be your first channel. Our guide to checking app demand with Google Trends and Keyword Planner shows how to read both tools.

For mobile apps, Apple's guide to App Store search explains how the store matches queries to apps. Store search can be strong in busy categories and quiet in niche ones.

Run three tests with one metric

Each test needs a budget, a time limit and a target, written before it starts.

  1. Pick three channels from your shortlist.
  2. Set a budget for each, in money or hours, such as 10 hours or a small ad spend.
  3. Choose one metric for all three: cost per activated user.
  4. Tag every link with UTM parameters so you know the source.
  5. Run each test for one to two weeks.
  6. Compare the cost per activated user and the quality of the conversations.

An activated user is someone who did the core job once. For JobBoard, that's a firm that added its first job and assigned it to a plumber. Clicks and sign-ups are easy to buy and tell you little.

Read the result as a JobBoard founder would

Suppose the tests end like this. Direct outreach: 40 calls, 6 activated firms, 10 hours. Search ads: 300 clicks, 4 activated firms, £300 spent, 10 hours. Supplier partnership: one meeting, no result yet.

Outreach wins on activation per hour, and each call taught something about the product. Search ads work but cost more per firm. The partnership needs more time and stays on a slow burner.

So outreach becomes the focus channel for 90 days. That doesn't mean ignoring the rest forever. It means one channel gets most of the effort until it stops improving.

Do things that don't scale, on purpose

Early channels often look unscalable: calling firms one by one, visiting a trade counter, answering every forum question by hand. Paul Graham's essay Do things that don't scale argues that this is how most startups start.

Manual channels teach you what to say, who buys, and what they need. Those lessons make the scalable channels (ads, SEO, partners) work better later.

Let AI help with the busywork

AI assistants save time inside each channel. They can build a list of local plumbing firms from public directories, draft five versions of an outreach message, or write ad headlines from your customers' words. Keep the human parts human. Read each message before it goes out, and follow each community's rules on self-promotion.

Measure your channel tests in GrowApps

A GrowApps app can send events to your own PostHog or Google Analytics account, so each channel's activated users show up in a tool you control (sign-in, data, AI and payments). For web apps, a findability checklist runs at every publish and you can connect Google Search Console, which helps if SEO is on your shortlist.

List your shortlist in ten minutes

Ask three customers how they found the last tool they started using. Write each answer as a channel. Pick the three most common and write one test for each, with a budget, an end date and the activation event you will count. Then plan the first week with the launch checklist for your first 100 users.

Frequently asked questions

Should I post on social media from day one?

Only if your customers use it to find tools. Posting builds an audience slowly. For most small business apps, direct outreach or search gets the first users faster. Social posts can support the main channel later.

How long before I give up on a channel?

Give a test the time and budget you set at the start. If it misses its target clearly, stop. If it is close, change one thing (the message or the audience) and run it once more.

Sources and further reading