Most wasted Search spend starts as a structural mix. The brand name, a specific SKU, a vague problem, and a competitor's trademark sit in one campaign because they all "relate to the store." They do not relate to each other as buying jobs. They need different pages, different negatives, and different rules for what good looks like.

Use four buckets, and name them the same way every time so a new ad group cannot hide between them.

The four buckets

  • Own brand. Queries that include your name. This budget defends demand you already created. It should look efficient, and it should not be asked to prospect.
  • Bottom-of-funnel product fit. Model names, sizes, materials, and collection nouns a person uses when they are close to choosing. These want the product or collection page, not the homepage.
  • Problem and use-case tests. Queries that describe a situation rather than your SKU: a climate, a room, a constraint, a job to be done. These are allowed to be tests. They need a page that answers the situation and a spend limit, because the intent is less explicit.
  • Competitors. Other brands and direct substitutes. This is optional. It is its own budget with its own creative, and it does not get to borrow headlines from the brand campaign or print the other company's name just because the query did.

If you cannot say which bucket a keyword is in, it is not ready to be a keyword. "Best" queries, gift queries, and category queries usually belong in product fit or in a problem test, and the landing page decides which. A gift query with no recipient attribute on the page is not a strategy yet.

Why the split has to happen before the keywords

Intent is not a label you paste on after the keywords exist. It decides the page, the negative list, and whether the query is allowed to share a budget with your brand. Build the buckets when you scan the store, then fill them with phrases you discovered from models and problems and verified in Keyword Planner.

A test bucket that shares a budget with proven product terms will either starve or, worse, spend when you are not looking because its CPC is cheaper. Give the test its own cap. End the test on a date. Do not let "problem / use case" become a permanent junk drawer for every keyword that did not fit.

Competitor buckets fail when they are actually brand queries in disguise, or when the ad promises a comparison the page does not contain. If you do not have a page that explains why a shopper would switch, do not buy the query. Read brand defense versus conquesting before that campaign exists.

Naming is the control

Use a consistent name: brand, product, problem, competitor, then the collection. When someone exports the account, the name should tell them the intent without opening the keyword list. Inconsistent names are how a fifth bucket appears, usually called "general" or "test 2," which is the original mix returning under a new label.

Review the structure when you add a collection. New products join an existing bucket. They do not justify a new campaign type by default.

How Kopa builds the four buckets

In Kopa, Smart Campaigns splits the draft into those four buckets: own brand, bottom-of-funnel product fit, problem and use-case tests, and competitors. Each bucket can carry its own budget logic, and brand terms stay out of the other three. You approve the structure before it publishes.

What good looks like

  • Every Search keyword sits in brand, product fit, problem test, or competitor.
  • Each bucket has its own budget and its own definition of success.
  • Problem queries are capped tests with a page that answers the problem.
  • Competitor campaigns are opt-in, with no competitor names written into your headlines.
  • Campaign names state the bucket, so a mixed "general" campaign cannot reappear by habit.