Google Ads account structure that concentrates spend.

Most underperforming Google Ads accounts do not have a bidding problem or a creative problem. They have a plumbing problem.

Budget is spread evenly across everything, which means a meaningful share of it is buying clicks from people who were never close to purchasing. No individual decision caused it. It accumulated, one reasonable-seeming campaign at a time, over about two years.

This is the first thing we rebuild, and it is usually why the first result arrives inside a fortnight.

What Google Ads account structure actually means

Account structure is how campaigns, ad groups, and budgets are organised, and it determines where your money goes before any bid is placed.

The hierarchy is fixed: account, then campaigns, then ad groups, then keywords and ads. Budget is set at campaign level. That single fact is the reason structure matters more than almost anything else you can adjust — the campaign is the unit at which you control spend, so campaigns are how you decide what gets funded.

Organise by product category and you have decided to fund each category roughly equally, whether or not each category deserves it.

The mistake almost every account makes

Accounts get built around the shape of the catalogue. One campaign per product line, sometimes one per collection page. It feels tidy and it maps cleanly to how the business thinks about itself.

The customer does not shop that way, and neither does the auction.

Some searches come from people ready to buy right now. Others come from people three months out doing research. Others come from people who will never buy anything. Those three groups sit inside the same product category, and a category-shaped account funds them identically.

The second failure is fragmentation. Splitting into many tightly themed campaigns to gain control produces a set of campaigns each too small to gather enough conversion data. Smart Bidding needs volume to work. Starve it and it never exits the learning phase, so you get the volatility of automation with none of the benefit.

Structure around intent instead

Group by how close someone is to buying and what they are worth when they do.

High intent. Specific product searches, model numbers, “buy” and “price” modifiers, competitor comparisons. These convert best and deserve the most budget and the tightest control.

Mid intent. Category and problem searches. People know what they want but not from whom. Real potential, longer path, so measure them on assisted conversions rather than last click.

Research. Informational queries. Occasionally worth funding for retargeting pools, frequently not worth funding at all. This is where most wasted spend lives.

Brand. Always separate. Branded search converts far better and far cheaper than everything else, and blending it into other campaigns makes broken campaigns look acceptable. If your account shows 4X ROAS and 60% of conversions are branded, your actual acquisition performance may be closer to 1.5X. Separating brand is the fastest way to find out what you are really paying for a new customer.

Consolidate before you split

The instinct when performance drops is to add granularity. It is nearly always the wrong direction.

The working rule: if a campaign is not producing around 30 conversions a month, it is probably too small to optimise. Merge it. You lose some theoretical control and gain something worth much more, which is a bidding algorithm with enough data to make decisions.

Most accounts we take over are running roughly three times the campaigns they need.

Budget follows evidence, in stages

Once structure is right, budget moves in tiers rather than in one jump.

Raise spend on a proven campaign by 20–30%, hold for a week, and check efficiency held. If it did, raise again. If it did not, you have found the ceiling for that campaign at that structure.

Slower than flooding a winner, and it is why Indosole reached 5X at three months instead of collapsing back to break-even in week four. That collapse is the usual outcome of scaling fast, and it gets misread as the campaign “burning out” when what actually happened is that spend outran the intent available at that efficiency.

The half that lives outside the ad account

A restructure improves how efficiently you buy attention. It cannot fix what happens after the click.

ROAS is roughly traffic quality × conversion rate × average order value ÷ cost per click. Three of those four have nothing to do with your ad account. You can hold flawless structure and still return 1.5X because the site converts at 0.8% and the average order is $60.

Nocs Provisions is the clearest illustration. Their account returned 0.5X, which looks unambiguously like a media buying failure. But the site converted at 0.8%, and at that rate no bidding strategy makes the maths work. Conversion went to 1.5% first — an 87% lift — and only then did the rebuilt ad structure take ROAS from 0.5X to 4X.

Two months spent optimising bids would have produced a beautifully optimised account that still lost money.

Doing the restructure

1. Audit where the money actually went. Sort campaigns by spend and by conversions. The gap between those two orderings is your problem, stated plainly.
2. Pull brand out. Immediately, before anything else. You cannot assess the rest until branded performance stops flattering it.
3. Map searches to intent tiers. Use the search terms report, not the keyword list. What people typed matters more than what you bid on.
4. Merge anything starved. Under ~30 conversions a month, it joins something larger.
5. Rebuild in intent tiers, high intent funded first.
6. Fix the landing destinations before increasing spend, because that is what decides what a click is worth.
7. Scale in stages, each tier proving it holds before the next unlocks.

Expect one to two weeks of learning-phase noise. Judge nothing inside that window.

What to expect

Indosole went 1.5X to 3X in two weeks on structural work alone, before any new creative existed, then 5X at three months once staged scaling ran.

That is a fairly typical shape when structure was the actual constraint. It is a fast result because you are not waiting on anything to warm up. You are redirecting money that was already being spent toward the searches that were always going to convert.

Which is also why it is worth doing first. Everything downstream gets judged more honestly once the plumbing is right.