Key takeaways
- Judge each search term against the keywords you chose to bid on, before you look at its conversions.
- Most B2B SaaS terms never convert within the review window, so conversion-led reviews miss slow waste.
- Sort terms into five buckets and act on each bucket differently.
- Add negatives at the right level (account, campaign or ad group) and match type, or you block good traffic.
Search term reports are where most wasted spend in a B2B SaaS account hides. The common way to review them is to sort by cost, look at conversions, and exclude anything expensive that did not convert. That catches the obvious waste, but it misses most of it.
The reason is timing. A B2B software buyer might click an ad, leave, come back through a branded search two weeks later and book a demo a month after that. In a seven-day or thirty-day report, most relevant terms show zero conversions. So do most irrelevant ones. Conversion data alone cannot tell them apart.
A better starting question is simpler: does this search term match what we chose to bid on?
Start from your active keywords
Your active keyword list is a record of the intent you decided to pay for. Every search term that triggered an ad can be compared against it. If you bid on "contract management software" and the ad showed for "contract management jobs", the mismatch is clear before you look at a single metric.
Export two reports for the same date range:
- Active keywords with campaign and ad group
- Search terms with campaign, ad group, clicks, cost and conversions
Then review each search term against the keywords in its own ad group.
Sort every term into one of five buckets
| Bucket | What it means | Action |
|---|---|---|
| On target | Matches the ad group's keywords and the product | Keep. Consider adding as an exact keyword if volume is meaningful. |
| Wrong ad group | Relevant to the business, but a better fit elsewhere | Add as a keyword in the right ad group and a negative in the current one. |
| Off target | Unrelated to the product or to buying intent | Add a negative at the right level. |
| Research intent | Relevant topic, early stage (definitions, templates, "what is") | Decide per account. Often a negative for demo campaigns, a keyword for content campaigns. |
| Competitor or brand | Mentions a competitor or your own brand | Route to the dedicated competitor or brand campaign. |
Conversions come in after sorting, as a tiebreaker. A converting term in the off-target bucket is worth a closer look, since it might reveal a use case you have not considered. A non-converting term in the on-target bucket usually just needs more time.
Place negatives at the right level
Negatives are powerful and easy to overdo. A few rules keep them safe:
- Account level for themes that never make sense: careers, salary, torrent, crack. Avoid the bare word "free" (and its local-language versions): it also blocks "free trial" searches.
- Campaign level for terms that belong in another campaign, such as competitor names in a generic campaign.
- Ad group level for routing between ad groups inside the same campaign.
- Phrase match for themes, exact match for single queries you want to block without affecting close variants.
Before adding a negative, check it against your active keywords. A negative that matches one of your own keywords silently switches that keyword off.
A weekly routine
- Pull active keywords and search terms for the last seven days.
- Bucket new terms, starting with the highest cost.
- Add negatives and new keywords in one bulk upload.
- Note any new themes, like a new use case or a competitor showing up more often, for the monthly strategy review.
The first review of an account takes a few hours. After that, a weekly pass usually takes under thirty minutes, and most of the sorting can be automated so a person only reviews the edge cases.
Where automation helps
Comparing thousands of search terms against a keyword list is repetitive work that software does well. We run the matching and first-pass bucketing automatically, then a strategist reviews the off-target and wrong-ad-group lists before anything is uploaded. The judgement stays with a person. The checking does not have to.
Questions
How often should we review search terms?
Weekly for accounts spending more than a few thousand euro per month, and at least monthly for smaller accounts. Broad match and Performance Max need closer attention than exact match.
Should we add a negative for every term that did not convert?
No. In B2B SaaS most relevant terms do not convert within a week or even a month. Remove terms that are off-target, and give relevant terms enough clicks to prove themselves.
Which match type should negative keywords use?
Use phrase match for clear off-topic themes such as jobs, crack or tutorial, and exact match when you want to block one specific query without affecting close variants.