Non-brand search is where new customers come from, and where small budgets disappear fastest. Here's how to set up a test with a hard ceiling, a stop rule and a clear answer at the end.
Brand searches are easy. People who search for your name were looking for you anyway, and they're cheap to win. Non-brand searches, for what you sell rather than who you are, are where new customers come from. They're also where a small budget can vanish in a week without a single sale.
You don't have to choose between not trying and burning money. You can run non-brand as a test, with a fixed amount you're willing to lose and a clear rule for when to stop.
The short version
- Decide up front how much the test may cost, and what result would make it worth continuing.
- Put non-brand in its own campaign with its own budget, so brand can't make it look better than it is.
- Start narrow: a handful of exact and phrase keywords for your best-selling products or services.
- Set a stop rule, for example: no sale after spending three times what a sale is worth to you means pause.
- Judge the test against your break-even, not against your brand campaign.
Before you start: what is a sale worth?
Work out what you can pay for a sale and still make money. For a shop, that's your margin on an average order. If you prefer ROAS, work out your break-even ROAS: what an order has to bring back for each euro of ad spend before it pays for itself. See What is a good Google Ads ROAS for my business?
That number is your yardstick for the whole test.
Set up the test
A separate campaign. Brand and non-brand behave very differently. In one campaign, brand's cheap sales hide non-brand's expensive ones, and the bidding moves money towards brand. Keep them apart, and exclude your brand from the non-brand campaign with negative keywords.
A fixed daily budget. Pick a daily budget you can live with for the whole test period. Remember that Google can spend up to twice the daily budget on a single day, but evens it out over the month.
Few, specific keywords. Start with exact and phrase match for searches that clearly mean someone wants to buy: product names, "buy", "price", the service plus your town. Leave broad match for later.
A cap on click prices, if you start on clicks. A new campaign has no conversions to learn from, so many start on Maximize clicks. Then set a maximum CPC bid limit, so one expensive click can't eat a day's budget. Move to bidding for conversions once the campaign has some.
The search terms report. Check it twice a week, and block searches that don't fit.
The stop rule
Google Ads has no setting that pauses a campaign after a certain spend without a sale. You have to keep the rule yourself, or with an automated rule or a script.
A simple rule: if the campaign has spent three times what a sale is worth to you without a single sale, pause it and look at the search terms before you try again. With sales coming in, judge the cost per sale or the ROAS after four to six weeks, not after four days.
Reading the result
- Sales at or below your break-even: keep it, and widen it slowly with more keywords.
- Some sales, but too expensive: look at which keywords and searches brought them, keep those, and cut the rest.
- No sales, but good clicks: check the landing page and the price before you blame the keywords.
- No sales and poor searches: stop. Not every business can win non-brand search at today's prices, and that's a useful answer too.
How Bavian handles it
Bavian splits your total daily budget across your campaigns, and moves brand searches into a campaign of their own once there are enough of them. You can fix a campaign's daily budget yourself, and Bavian keeps it and works around it.
