Report unclassified search-term spend without inventing the missing queries

Recognize the evidence limit

An owner may see campaign spend above the sum of visible search terms and assume the account is hiding avoidable waste. First check the report definitions. Google omits some low-activity queries for privacy, and search-term insights can group requests differently. A missing query is not proof of irrelevant intent, nor is it proof of useful intent. The defensible report distinguishes what was classified from what could not be classified.

Source checks: About the search terms report · Understand your conversion tracking data

Align the two exports before subtracting

Select identical dates, campaigns, networks, and relevant filters. Check that both reports use cost rather than clicks or conversion value, and avoid mixing an account total with a single campaign's terms. Note the export time so later processing changes are not mistaken for an arithmetic error. If a conversion column is included, use the same action and counting definition. Reconciliation starts with comparable inputs, not a blanket assumption about every difference.

Create three visible-intent buckets

Classify known suitable service requests, known unsuitable requests, and ambiguous visible requests. Keep the undisclosed remainder as a fourth row. Use written definitions and a few examples so another reviewer can reproduce the classification. Suitable intent does not guarantee a booked job; unsuitable intent needs a specific service mismatch. Do not move a query into the waste bucket only because its limited sample produced no conversion.

Calculate two different percentages clearly

Known unsuitable spend divided by visible spend describes the visible sample. Known unsuitable spend divided by campaign spend describes the confirmed unsuitable portion of the whole campaign. These are not interchangeable percentages. The unclassified remainder has no observed intent label. State the denominator next to each figure, and avoid applying the visible sample's waste rate to the hidden portion unless you have a justified analysis and explain its uncertainty.

Use grouped insights as context, not recovered transcripts

Where search-term insights are available, inspect themes for potential service or message issues. A theme can suggest a question to investigate, but it does not reconstruct every withheld query or customer outcome. Check the time range and grouping definitions. Preserve the evidence you used in the report note. This makes it possible to revise the classification later without pretending that a theme label was a complete list of actual searches.

Choose actions supported by the known portion

Add careful exclusions for clearly inappropriate visible intent, improve the destination for useful themes, and review actual inquiry quality. Keep unclassified spend in subsequent reports so the evidence coverage remains visible. If reporting coverage changes, explain that before comparing waste percentages across months. The useful decision is about a specific observed mismatch; a dramatic invented total usually makes the owner less able to choose the right correction.

Worked example: illustrative figures

For example, campaign spend is $500 and visible search terms total $350. Of that visible amount, $70 is clearly unsuitable, $245 is suitable, and $35 is uncertain. Unclassified spend is $500 − $350 = $150. Known unsuitable spend is 20% of visible spend ($70 ÷ $350) and 14% of total spend ($70 ÷ $500). Neither calculation says that all $150 unclassified spend was wasted. The report labels both denominators and uses the known unsuitable queries to plan narrow corrections.

Questions about this task