If you manage a Google Grant, you've probably had a stretch where the campaign just... stalls. Impressions dry up, clicks disappear, and the $10,000/month in free advertising sits mostly unused.

In my experience auditing Grant accounts, the underperformance almost always traces back to one of five issues. Here's what to look for and how to fix each one.

1. Tracking isn't set up – or isn't working

Google Grants requires accounts (created on or after April 22, 2019) to run on a conversion-based Smart Bidding strategy like Maximize Conversions or Maximize Conversion Values. Google's algorithm sets bids automatically to get you the most conversions (or value) within your budget, which takes the guesswork out of managing bids yourself.

The problem is that Smart Bidding only works if Google has conversion data to learn from. I've audited many Grant accounts that weren't running simply because conversion tracking was never set up, or it broke at some point and nobody noticed. With no conversions coming through, the algorithm has nothing to optimize toward which makes these campaigns stall out almost entirely – low impressions, few or no clicks, even with the full budget available.

To fix this, you need at least one meaningful conversion action set up. The conversion should represent a real action tied to your mission (like a completed donation, a volunteer sign-up, a newsletter subscription) rather than a passive one like a page view. You have two main paths:

One detail that trips people up: if you're tracking donations, make sure they're categorized correctly in Google Ads (as a "Purchase" conversion action) rather than lumped in as a generic conversion. Google uses this categorization for its own reporting on the program's impact, and miscategorized conversions can create compliance headaches later.

Aim for enough conversion volume to actually give the algorithm signal. Most guidance puts the useful threshold at around 15 to 30 conversions a month but honestly, the more the better. If your current conversion point doesn't generate that kind of volume, consider tracking a slightly earlier, higher-funnel action (like a newsletter signup) in addition to your primary goal.

2. Keywords are too broad or generic

Grant accounts can't bid on single-word keywords, with a short list of exceptions (your own brand terms, recognized medical conditions, and a few published exception terms). Keywords with dashes, periods, or special characters aren't treated as single-words but the intent behind the rule is the same: Google wants every keyword to clearly reflect what someone is searching for and how it connects to your mission.

This extends beyond single words. Overly generic phrases like "free videos," "things to do," & "today's news" get flagged for "unclear relevance," even if they're technically two or three words. And any keyword sitting at a Quality Score of 1 or 2 has to be paused; leaving it active risks a policy violation.

The fix is specificity. Use the Keyword Planner to find variations that are closely tied to your organization. Think brand-term variations and long-tail phrases that clearly signal intent, rather than the broad terms your mission might use in casual conversation. "Museums in [city]" works; "museum" on its own doesn't.

3. Outdated account structure

If a campaign used to perform well but has quietly declined over the past few years, it's often a structural problem rather than a budget or targeting one. Highly granular ad groups, including single-keyword ad groups (SKAGs), which used to be a common PPC best practice, generally don't perform as well anymore. Smart Bidding algorithms need volume and data to optimize effectively, and splitting keywords into very small ad groups starves each of the signals the algorithms need.

In my experience, slightly larger, tightly themed ad groups give the algorithm more to work with while still keeping the ads relevant to what's being searched. You should strive for at least 15 to 20 closely related keywords per ad group. Consider removing low-quality score keywords to keep the ad group high quality overall.

4. Ad strength is poor

Google scores your responsive search ads with an Ad Strength rating (Poor, Average, Good, Excellent). A poor score generally means your ads are competing at a real disadvantage for impressions and clicks.

When a Grant isn’t performing well, one of the key culprits is a poor or even average Ad Strength. The reality is that you really need a “Good” or higher ad strength to start getting traction on non-branded keywords.

To improve it:

Aim for a "Good" Ad Strength rating as a baseline; "Excellent" is the ceiling but not always necessary to get solid performance.

5. Targeting is too broad or too narrow

There are typically two issues I see when a Grant holder sets up geographic targeting. Its either too broad, targeting the “All Countries” or too narrow, set to a radius of 30 mi around a city. Neither of them are good options.

It’s extremely rare to find a Grant organization that is truly wanting to target the entire world. To do so well would require multiple languages, support across the globe and certainly more than just $10,000 a month. Instead target exactly what you need. For most US based organizations that will be either the country or region that they operate in.

Why not city? You might be thinking that city based or radius based targeting is right because it’s highly targeted. The problem is that in practice, city & radius based targeting isn’t as accurate as one might think. I’ve found that broadening from a radius based target to a Nielsen DMA region works significantly better.

Remember: Google is Data Hungry

Everything above points to the same underlying principle: Google's Smart Bidding and Quality Score systems are data-hungry. The more accurate signal you give them – clean conversion tracking, well-structured ad groups, varied ad copy, tight keyword relevance – the better your account performs, and the more of your monthly grant you'll actually put to use.

If your account is stuck, it's worth working through these five points in order. Tracking issues alone account for a large share of underperforming Grant accounts I see, so that's always the first place to check.