A SaaS vendor measured its brand-search spend honestly and moved half the budget

Representative examplesEvery case study in this library is an illustrative composite of the kind of engagement we deliver — written to show our method and standards, not to name clients.

Where they started

Growth at the B2B SaaS vendor had made the brand name a shorthand in its niche, and branded search volume reflects it — prospects arrive at Google already primed, typing the product's name. Sales runs demo-led, the CFO reviews marketing monthly, and competitor ads against the brand had multiplied over a year. Brand search was the account's untouchable line: near-perfect reported ROAS, defended by the logic that competitors were bidding on it, growing without a measurement in sight.

What it was costing

The account's best 'performing' campaign was brand search — near-100% ROAS, uncontested, beloved. Competitors had started bidding on the brand name, so 'defensive' spend had doubled. Nobody could say how much was defense and how much was paying for customers who were coming anyway.

What they could see

• Brand-search spend had doubled across the year with no change in the branded traffic the site's analytics reported.
• The campaign's near-perfect ROAS made it untouchable — any suggestion to examine it died against the number itself.
• Competitor ads sat above the organic result on brand queries most mornings, and the response had always been to bid higher.
• The CFO began asking what portion of the line was defense versus customers who were coming anyway; nobody had an answer.

The constraints we worked inside

• Competitor bidding on the brand was real and growing — some defense had merit.
• The CFO wanted the budget line justified; faith-based ROAS would not survive the meeting.
• Pausing brand search entirely was too risky — a competitor would fill the space.

What had been tried before

Brand bids were raised reactively whenever competitor ads appeared on the brand name.

Reactive raising pays competitor-auction prices across every brand query, including the many where organic already wins — defense priced like panic.

Brand search was paused globally for a single day as a test.

One day mixes weekday patterns, hands competitors a free window, and produces numbers too coarse to separate organic capture from paid defense.

A spreadsheet estimate of organic-versus-paid overlap was prepared for the CFO meeting.

Assumed capture rates dressed as math; the meeting needed measured evidence, and modeled opinion collapsed the first time someone asked where a number came from.

What we proposed

We proposed measuring the brand line like an insurance policy. A two-week geo holdout — ads paused in matched regions — would separate organic capture from paid defense. Where the holdout showed no loss, organic-only brand queries would come out of paid; where competitors were live, bids would reset to a cost-per-defense basis on competitor-triggered queries. The budget line would then report competitor impression share and defensive win rate instead of a near-perfect ROAS nobody trusted once examined.

Just as important is what we ruled out, and why:

• Raising bids on all brand queries — Blanket raises pay competitor-auction prices on every brand search, including the majority where organic already captures the click.
• Pausing brand search entirely — The constraint ruled this out upfront; competitors were actively bidding and an empty slot on the brand name was a real risk to demo-day revenue.
• Trademark complaints as the remedy — Complaints take weeks, apply unevenly across regions and reseller exceptions, and do nothing about the spend already committed while waiting.

How the work ran

Delivered by the growth pod — paid search specialist over 5 weeks, with working increments reviewed with the client every week.

The stack, and the reasoning

Google Ads

Brand defense happens in this auction whether we like it or not — the work was measuring it, and the levers live in the account.

Geo-holdout test

Pausing in matched regions separated organic capture from paid defense, the one question the account's ROAS numbers could never answer.

Auction insights

Competitor impression share on brand terms became the defense metric — it prices the threat without needing competitor spend estimates.

Query-level attribution

Splitting competitor-triggered from organic-only brand queries let the line pay for defense only where defense is real.

What went wrong

How we worked together

Cadence

A weekly thirty-minute call with the marketing lead during the holdout, then fortnightly; the CFO joined two sessions — one before the test, one after the result.

Client side

The marketing lead owned the account; the sales lead received a daily branded-traffic summary during holdout weeks and returned one question, once.

Decisions

The CFO approved the decision rule in the pre-test session; the result then executed itself — the meeting after was shorter than the one before.

They provided

Analytics access for branded-organic monitoring, sales-team buy-in for the holdout regions, and the CFO's explicit tolerance for two weeks of unoptimized brand spend.

What changed

The headline: brand-search spend after the holdout, with branded-organic sessions unchanged (within 2%) — −52%, read from Holdout results + Search Console. A second check: competitor impression share on brand terms (defense working, cheaper) at 94% → 71%.

The brand line gets discussed as insurance, with a premium and a coverage level, instead of being defended by a ROAS figure nobody believed. The CFO's monthly review runs faster because the number presented is the one he asked for. Sales stopped noticing brand-ad gaps because the defense now concentrates exactly where competitors actually bid. The marketing lead describes the account as spending where the fight is — and the spend that vanished was never doing defensive work at all.

The result was read from Holdout results + Search Console against the pre-engagement baseline over the stated window, with a guardrail check on competitor impression share on brand terms (defense working, cheaper). Where platform-reported numbers and business outcomes differ, this record says which layer it is quoting.

What they own now

• The cost-per-defense bidding structure on competitor-triggered brand queries.
• The geo-holdout methodology and results, archived for the next incrementality question.
• The defense reporting format — competitor impression share and win rate — for the CFO pack.
• Query-level attribution setup separating competitor-triggered from organic-only brand traffic.
• Auction-insights monitoring routine with thresholds for when defense bids adjust.