How Grammarly got its users

Grammarly's ad spend only clears because the install is the activation event. The extension sits in Chrome, Safari, Firefox and Edge and fires underlines inside Gmail, Docs and any text field, unprompted, every day. That is a surface, not a visit. Before you buy one click, confirm your product has a place it shows up on its own. If people have to remember to open you, ads buy trials that quietly die in a week.

done whenIn your own analytics: what share of week-one users trigger the product again in week four with no email from you? If you cannot separate "came back alone" from "came back after a nudge", you are not ready to spend.

The ads do not land on /signup. They land on grammarly.com/grammar-check and /paraphrasing-tool, whose page title reads "Paraphrase Instantly (No Sign-up)". You paste text, underlines appear, you accept a fix. Only after that does "Sign up to keep going" show, offering tone and clarity insights. Build that page before you build the campaign: one paste box, visible output, the wall placed after the value. Every paid click goes there.

done whenOpen your own ad destination on a phone, logged out. Count the seconds until a stranger sees output made from their own input. Past ten seconds, or an email field first, and you paid for that click twice.

Platform: YouTube. Squero's published Grammarly case study gives the brief away: the product demo is treated as the hero character, and the absurd framing (commentators calling desk work like a sport, the whole set sunk underwater) exists only to hold the first five seconds around it. With no budget, skip the flooded set. Screen-record one real correction landing inside a live Gmail compose window, no voiceover, and let the shot run long.

done whenWatch your own cut with the sound off and freeze it at five seconds. If a stranger cannot say what the product does from that frame alone, the rest of the runtime is paid for and unwatched.

Second surface: paid search on the job query, not the category. Grammarly keeps one page per job with the query as the headline, "Free Grammar Checker" at /grammar-check, a rewording tool at /ai/ai-writing-tools/rewording-tool. Someone typing "free grammar checker" wants a box, not a comparison of writing assistants. Pick your two narrowest jobs, give each its own page and its own ad group, and match the headline to the query word for word.

done whenList every query you would bid on. For each, do you have a page whose headline repeats those exact words and whose first screen does the job? No page, no bid. Pointing them at your homepage burns the budget.

Then check the meter. Grammarly Free hands out 100 AI prompts a month, Pro raises that to 2,000 and opens as a 7-day trial with a reminder emailed two days before it ends. The free install is worth paying for because a counter turns it into revenue with no salesperson attached. If your free tier has no unit that runs out and no reason to hit the ceiling, ads buy you users and nothing else. Set the counter before you open the ad account.

done whenName the unit your free tier meters and the number where it stops. If you cannot state both in one sentence, or nobody has hit that ceiling yet without ads, spending only grows a free user base.