Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Google Ads › Expert

Adding Performance Max for Qualified Leads for a Removals Firm

You decide whether an invented removals firm, which wants Performance Max because a competitor runs it, is ready to give the campaign a signal worth optimising to. You finish with a readiness gate for the qualified outcome, the goal the campaign bids towards, asset groups chosen by service and value, the controls and division of work with Search, an account-level reading plan and a stopping rule agreed before launch.
ExerciseExpert levelModule 3: Account architecture for scale10 minUpdated

Why this matters

The advanced Google Ads level separated brand demand from a retailer's feed-based Performance Max and fed bidding better conversion signals. A lead business meets the same campaign type with nothing to anchor it: no feed, no prices, no order value arriving with the conversion. Performance Max optimises to whatever the account calls a conversion, so the decision that matters is made before launch, in the job records rather than the campaign settings.

"A competitor does it" is not a case. A case names the demand Search is not reaching, the qualified outcome the campaign will learn from, the controls that keep it honest and the evidence that will end it if it adds no customers. The guide to Performance Max for lead generation sets out each part; here you apply them for an owner who wants the campaign live before the summer.

The removals firm and its proposal

Dunvorral Removals, an invented house and office removals firm with a storage depot, sells local house moves, long-distance moves, office removals and storage, and runs a small online shop for packing boxes. Enquiries arrive by the quote form and by phone, an estimator surveys house and office moves in person or by video, and the job system records each enquiry's service, its survey and whether the job was booked. All paid traffic runs through Search today.

Illustrative example — not a real result

Dunvorral Removals: the account, the job records and the agency's proposal

Google Ads, last quarter. Primary actions: quote form submissions and calls from ads, both counted every time.

CampaignTargeting and biddingSpendEnquiries recordedSearch impression share lost to budget / rank
BrandExact match, maximise clicks£1,3002400% / 6%
House removalsExact and phrase, maximise conversions£13,5004602% / 19%
Office removalsExact and phrase, maximise conversions£4,500540% / 24%
StoragePhrase, maximise conversions£2,8001500% / 15%

Job system, enquiries from paid search including brand, same quarter:

ServiceEnquiriesSurveys bookedJobs bookedAverage margin per jobUsual timing
Local house moves420202107£540Survey within a week, job booked two to six weeks later
Long-distance moves1105226£580As local moves
Office removals48216£3,300Site survey within two weeks, contract one to three months later
Storage150Not surveyed63 units let£400 over the average stayUnit let within a week

The state of the signal:

  • The quote form was rebuilt seven weeks ago. Since then it stores the click identifier and the consent choice with each web enquiry; 91% of web enquiries from paid search carry an identifier. Before the rebuild, none did.
  • 41% of enquiries arrive by phone, and the job system stores no identifier for them.
  • "Survey booked" is a stage the office team sets by hand. A check of 60 recent surveys found 8 whose stage date had been overwritten when the survey was rescheduled.
  • No offline upload has ever run. The office manager has offered to own it.
  • Last year, June to August held 41% of the year's house moves. The owner will pay at most £180 in advertising for a booked house move.

The agency's proposal, written after the owner saw a competitor's video and display ads: one Performance Max campaign with one asset group for every service and the homepage as its final URL; bidding on all enquiries with a target CPA of £30; final URL expansion on; no brand exclusion "so it can find the most volume"; £5,000 a month moved from the House removals campaign; launch in three weeks. The site has pages for each service, the packing shop, a careers page and a "track my move" customer log-in.

A campaign with no feed

Without a product feed there are no listing groups and no prices to steer by. Everything the campaign knows about relevance and worth then comes from the asset groups, the pages they send people to and the goal it bids on, so quality is decided by what the goal counts. That is why the guide puts the qualified outcome first, returned by an offline conversion import or enhanced conversions for leads, with the stage chosen as a trade between truth and volume. Two of its rules then decide the structure: asset groups share their campaign's budget and target, so a service whose qualified lead is worth a very different amount needs a campaign of its own, and a thin service belongs inside a larger campaign rather than in one of its own, so where no larger campaign shares its value it is left to Search. Lead goals never share a campaign with sales goals, and whatever the campaign does not own is kept out by controls.

Task

  1. Decide whether the firm is ready to launch, and if not, write the gate it must pass first and who owns each part.
  2. Choose the outcome the campaign bids towards, say whether that signal is clean enough to launch on, and what the existing Search campaigns keep.
  3. Choose the services the campaign covers and its asset groups, and give a reason for each service left out.
  4. Set the controls that protect lead quality and brand traffic, and say where each one is applied.
  5. Divide the work between the new campaign and Search, including how it is funded.
  6. Write how the campaign is judged on qualified leads, against what ceiling and baseline, and the rule under which the firm scales it back or stops it.

Worked answer

Readiness. The account makes a narrower case than the agency's. House removals loses almost nothing to budget, so more money in Search buys little; what it loses to rank comes next, through a looser target and wider queries, which the guide to scaling Google Ads spend takes before any new campaign type. That step is taken and read in Search while the gate is built, and Performance Max is then a case only for the house-move demand Search still does not reach. The signal is not ready: identifiers go back only a few weeks, the survey stage date is unreliable, nobody has run an upload and phone enquiries cannot be matched. The gate: the job system's owner adds a field that keeps the first date a survey was booked; the office manager runs a weekly upload under the offline conversion upload SOP, with its consent filter, log and note to paid media; and launch waits until consecutive uploads track the job system's own count of surveys from identified web enquiries, whatever the calendar says.

What it bids towards. Choosing a stage and valuing it from close rate and margin is settled practice from the advanced level. Here it gives a booked house-move survey from a web enquiry, counted once per click, as the campaign's own primary goal, with the raw form secondary and booked jobs uploaded later as a secondary action carrying their margin. Once the gate holds, that signal is clean but partial: phone enquiries never match, so the campaign under-credits itself for callers, one more reason the read comes from the job system. The Search campaigns keep their goals and every-time counting until the launch has been read, so that one result has one change behind it.

Services and asset groups. The value split decides the structure. Local and long-distance surveys turn into jobs at a similar rate and are worth about the same, so one lead campaign for house moves holds an asset group for each, sending people to its own page, with creative and search themes for that service and audience signals from its past customers where consent allows. An office survey is worth several times a house-move survey, so a shared target would be wrong for both; office surveys are also too few for a campaign of their own, and with no larger campaign sharing their value, office removals stay in Search. Storage stays in Search because its qualified outcome is a let unit, not a survey. The packing shop is a sales goal and never enters a lead campaign.

Controls. A brand exclusion leaves the firm's name to the Brand campaign. Account-level negatives cover what every campaign should refuse: jobs, driver vacancies, van hire, free boxes and do-it-yourself searches. Office and storage terms are not added there: an account-level list reaches beyond this one campaign, and those services still sell through Search. They are kept out by URL exclusions, the absence of a matching asset group and campaign-level negatives where the current help page shows them available. Final URL expansion goes off, or stays on with exclusions for the careers page, the customer log-in, the shop and the office and storage pages. The quote form asks for the move date, property size and the areas moved from and to, and the tag fires on the confirmation page only. The Performance Max launch checklist is run at switch-on, its feed lines marked not applicable.

Beside Search. Following the comparison of Performance Max and Search campaigns, the brand and the house-move queries that book most surveys stay in Search on exact match, which takes priority when the query is identical. Search themes cover what the two pages under-describe and never copy the keyword list, and a weekly review of search term insights moves proven queries into Search as exact match. The new campaign gets new budget: cutting House removals to fund it would shrink Search by design, and the read would show a swap rather than an addition.

Judging and stopping. The read is at account level, from the job system: house-move surveys booked, the cost per survey booked, the share of house-move enquiries reaching a survey, and booked jobs once they catch up; campaign columns are for diagnosis. The ceiling comes from the owner's limit per booked move and the survey-to-job rate, worked with the CPA calculator. Each period is judged on the extra surveys the extra spend bought, against the weeks before launch carried forward by last year's change between the same two periods, so the season is not credited to the campaign. The firm's one service area is too small to split into matched regions for the Performance Max uplift test, so the result is reported as indicative, with brand Search impression share as the guardrail. The rule, signed before launch: the campaign pauses at once if the upload stops; after the settling period it gets one round of fixes if the extra surveys cost more than the ceiling or the survey share falls, and it stops if the next read is the same.

Illustrative example — not a real result

Dunvorral Removals: the arithmetic and the first read

The value split, per survey booked: local £540 × (107 ÷ 202) ≈ £286; long-distance £580 × (26 ÷ 52) = £290; office £3,300 × (6 ÷ 21) ≈ £943.

Ceiling per house-move survey: £180 × (133 ÷ 254) ≈ £94.

The gate was passed after six clean weekly uploads, the last two within 4% of the job system's count. The first read covers eight weeks once the settling period had passed:

MeasureLast year, eight weeks before the dateLast year, the same eight weeksEight weeks before launchEight weeks with Performance Max
Spend, all campaigns£12,400£12,900£13,600£17,700
Enquiries recorded in Google Ads505530560780
House-move surveys booked, job system140149158189
Share of house-move enquiries reaching a survey50%50%49%41%
Brand Search impression share93%93%92%91%

Performance Max reported 210 enquiries at about £19.50 each. Surveys rose by 31 against the weeks before launch, 14 long-distance and 17 local, but last year the same weeks brought 149 ÷ 140 ≈ 1.06 times the surveys of the weeks before them. Without Performance Max about 158 × 1.06 ≈ 168 were expected, so the extra £4,100 bought 189 − 168 = 21 extra surveys, about £195 each, double the £94 ceiling. Many of the new local enquiries were single-item moves from outside the service area. Under the signed rule the local asset group's page and form were tightened and single-item terms added as negatives; if the next read is still above the ceiling, the campaign stops and the firm stays on Search.

Common mistakes

  • Launching because a competitor did. The case lives in the account's headroom and the signal, not in a rival's ads.
  • Launching before the upload is a routine. A campaign bidding on raw forms learns to buy the cheapest enquiries.
  • One asset group for every service, on the homepage. Rejected enquiries cannot be traced to a service.
  • Crediting the season to the campaign. A launch into the summer rise looks like growth until last year's change is taken out.
  • Writing the stopping rule after the first read. A rule written once the result is known bends to fit it.

The capstone that closes this level, a Google Ads operating plan for a hot tub retailer, asks the same question of a retailer's survey leads beside its Shopping and Performance Max operation.

Pages this lesson applies

Terms used on this page

Also related

Related