The Google Ad Grant, and the rule that quietly kills most accounts

The Google Ad Grant gives an eligible non-profit up to 10,000 US dollars a month in free Google Search advertising. It is the largest budget most small organisations will ever have, it costs nothing, and a great many of the organisations that hold one are not spending it.

The reason is almost never the application. It is a performance rule that runs quietly in the background of every granted account and switches it off when the account stops clearing the bar.

What does the grant actually give you?

Text ads on Google Search, and only there. The grant does not cover the Display Network, YouTube, Shopping or anything on a partner site, so it buys you exactly one thing: the moment when somebody types a question your organisation can answer, and you appear.

Granted ads sit below paid ads in the auction, which matters less than it sounds. Most of the searches a non-profit should want are ones commercial advertisers are not bidding on at all.

What is the rule that closes accounts?

An Ad Grant account has to hold a click-through rate of at least five percent, measured across the account, every month. Two consecutive months below that and Google can deactivate the account.

Five percent is a demanding number in paid search. Ordinary commercial accounts often run below it and are perfectly healthy, because they are buying volume. A grant account rents a privilege, and the rent is paid in relevance.

The organisations that lose their grant are usually the ones that set it up once, pointed it at broad terms like "charity" or "donate", and left it alone. Broad terms attract clicks from people who wanted something else, the rate sags, and nobody notices until the account is off.

• Account click-through rate above five percent, every month · the usual cause of death

  Two months below and the account can be deactivated. This is the one that does the damage.

• Specific keywords rather than broad ones · the usual cause of death

  Generic single words attract the wrong clicks, and the wrong clicks are what pull the rate down.

• Conversion tracking that records something real · keep this true

  Without it you cannot use the smart bidding that lifts the rate, and you cannot prove the grant did anything.

• Geographic targeting set deliberately · keep this true

  A required setting, and one that stops you paying attention to clicks from places you do not serve.

• At least two ad groups per campaign, two ads per group, sitelinks in place · keep this true

  Structural requirements. Easy to satisfy once, easy to break during a website change.

Google updates the Ad Grant policy from time to time, so treat this as the shape of the requirement and check the current version before you act on any single number.

Why does the rule exist at all?

Because Google is giving away inventory that it could otherwise sell, and a bad ad occupies the same screen space as a good one. The five percent rule is Google's way of making sure free advertising is still advertising somebody wanted to see.

It is worth understanding that this makes the grant a fundamentally different instrument to a paid account. In a paid account you can buy your way through low relevance. In a granted account, relevance is the only currency you have.

How do you actually hold the rate?

By bidding on what people search when they already want what you do, rather than on what your organisation calls itself.

An organisation running a bursary programme should not be bidding on "education charity". It should be bidding on the questions a seventeen-year-old types at eleven at night: how to apply for a bursary, what documents you need, whether you can still apply without matric results. Those searches convert into the thing the organisation exists to do, and they convert into clicks because the ad is the answer.

That is the whole discipline. Write down the twenty questions a person asks in the week before they need you, and build the account around those, not around your mission statement.

What does a grant account not fix?

It does not fix a website that cannot survive the visit. A grant sends people to your pages at a volume you have probably never had, and if those pages do not tell a person what to do next, the traffic is a statistic rather than an outcome. I have written separately about the look-you-up test a funder runs before they agree to meet you, and the same test applies to a stranger arriving from a search.

It also does not fix measurement. If nothing on your site records an enquiry, a sign-up or a donation, you will end the year with ten thousand dollars a month of free advertising and no way to say what it did. That is a worse position than having no grant at all, because it is the position where you cannot defend the programme to your own board. The plain version of that problem is in what a conversion action is, and why yours might be lying.

Is it worth the work?

For most organisations, yes, and by a wide margin. There is no other channel where a small non-profit can reach ten thousand dollars a month of qualified attention without a fundraiser attached to it.

But the honest framing is that the grant is not free money. It is free inventory with a maintenance obligation, and the organisations that get value from it are the ones that treat it as a programme with an owner rather than a benefit that was switched on once.

If you would like the eligibility and the current account health checked before you commit anyone's time to it, that is the work I do for non-profits, and the check itself costs nothing.