How limits shape healthy keyword lists

The ceilings, plainly stated#

Each search results ad group holds up to 5,000 keywords and up to 5,000 negative keywords. Those numbers sound roomy until discovery runs across multiple locales, close variants multiply, and three years of launch guesses fossilise in place. Treat the ceiling as a design constraint from day one: every list should justify its size, and no ad group should need the full 5,000 to do its job.

Most healthy ad groups run in the dozens or low hundreds of live terms; thousands signal hoarding, not coverage. Our relevancy guide sets the fit bar every term must clear, and negatives cover the blocking half of the ceiling.

Building lists that fit#

Build lists by theme and job, not by dump: one intent per ad group, brand apart from category, discovery apart from control. Draft lists in a spreadsheet first, with columns for term, type, campaign home and reason, so review takes minutes instead of archaeology. Apple own keyword management help details the adding flow including spreadsheet upload for long lists.

Decide match type before upload, since type locks at save and bulk mistakes multiply the pause and re-add chore. Our exact match guide covers the lock, and broad match shows what the default gives you.

Bulk upload without tears#

Spreadsheet upload is the only sane route past a few dozen terms: prepare the template, separate terms cleanly, mark exact terms with brackets, and validate a small batch before committing the full list. Upload into the correct ad group the first time, because misplaced thousands are a special kind of Monday. Keep the source sheet as the list of record afterwards, versioned by date.

Reconcile the sheet against the account quarterly; drift between the two is where duplicates and zombies breed. Our harvesting guide shows how promoted terms should flow into both the account and the sheet together.

Pruning cadence that holds#

Prune on a fixed rhythm: monthly for negatives and dead positives on spend, quarterly for full list audits including zero impression terms. Pause what never served after fair testing, negate what repeatedly spends without converting, and delete nothing without a note, since paused history still teaches. A list that only grows is a list that is slowly lying to you.

Watch the negatives ceiling hardest, because junk arrives faster than winners and discovery in several locales fills quotas quietly. Apple own negative keyword help covers bulk routes for big cleanups. Our Search Match guide warns what unfenced automation does to list sizes.

Duplicates and overlap#

Duplicates across ad groups force your own terms to compete for the same queries, splitting history and muddying bids. Dedupe lists on upload, run overlap checks after every harvest promotion, and keep discovery fenced with exact negatives of everything controlled campaigns own. One term, one home, one history: that is the whole rule.

Overlap audits pair naturally with the quarterly prune, since both ask where each term lives and why. Our brand, category and competitor guide draws the fences that prevent most overlap before it starts.

When to split the ad group#

Split the ad group when themes fork, when one intent's bids clearly differ from its neighbours, or when the list approaches four figures and reviews take all afternoon. Splitting is cheap and reversible; suffering a 4,000 term group is expensive and endless. Name children for their intent so the next review starts oriented.

Start at the Apple Ads hub for the wider structure these groups plug into, and ask our team for a list audit when the ceilings start feeling close.

Frequently asked questions#