Skip to content
by CRANK 11
⌘K Free audit

Services Search and Shopping

Google Ads management · Search and Shopping

Someone is searching for exactly what you sell, right now. Someone else’s ad is there.

The people already searching for what you sell, caught at the moment they search: intent-sorted campaigns, a clean feed, tracking that counts once, and bids moved weekly against cost per result.

Search terms sorted by intent into three tiers, each with its own bid, its own ad and its own page. Shopping fed by a corrected catalogue. Negatives grown every week from the search-terms report.

Book a Profit Leak Strategy Session

30 minutes with Jack or Max. We read your account live and tell you where the money leaks.

A brass magnifying glass standing on black slate, a gold beam of light landing on a small storefront

Intent is the whole game

Three kinds of search. Most accounts bid on them the same way.

"Best running shoes for flat feet" and "buy brand X size 42" are not the same person. One is comparing, one is buying. An account that sends both to the home page with the same bid overpays for the browser and underpays for the buyer, then blames the channel.

Search is the one channel where the customer tells you what they want in their own words. The work is to listen: sort the terms by intent, write an ad that answers each tier, send each tier to the page that finishes the job, and grow the negatives every week so the budget stops leaking to the wrong words.

Shopping is the same discipline applied to a feed: titles rewritten for how people search, not how the warehouse labels things, and a structure that lets the best products win the auction on their own.

Three search intent tiers and what we bid, write and build for each Problem, for a query like "best running shoes for flat feet": bid low, the ad is a guide, the page is a comparison. Compare, "brand X vs brand Y": bid mid, the ad is proof, the page is the product. Buy, "buy brand X size 42": bid high, the ad is the offer, the page is the checkout. BIDADPAGEProblem“best running shoes for flat feet”lowguidecomparisonCompare“brand X vs brand Y”midproofproductBuy“buy brand X size 42”highoffercheckoutThree search intent tiers and what we bid, write and build for each Problem, for a query like "best running shoes for flat feet": bid low, the ad is a guide, the page is a comparison. Compare, "brand X vs brand Y": bid mid, the ad is proof, the page is the product. Buy, "buy brand X size 42": bid high, the ad is the offer, the page is the checkout. Problem“best running shoes for flat feet”BIDlowADguidePAGEcomparisonCompare“brand X vs brand Y”BIDmidADproofPAGEproductBuy“buy brand X size 42”BIDhighADofferPAGEcheckout
The three intent tiers we sort every account into, with what we bid, write and build for each. This one is a method, not a measurement.

How it runs

Audit, build, tune every week.

  1. Audit

    24 hours

    Where the spend goes now, which terms leak it, what the auction costs for your words, what the feed is missing.

  2. Build

    Week one

    Intent tiers, campaign structure, the feed corrected, conversion tracking that counts once, first campaigns live.

  3. Grow

    Every week

    The search-terms report read line by line. Negatives added, winners promoted to exact match, new terms tested.

  4. Scale

    Every third day

    Bids and budget moved against cost per result and, for stores, margin per product. Every move in the report.

We never optimise to clicks. Every bid answers one question: what did that click cost per result, and what is a result worth to you.

Deliverables

What lands on your desk.

A structure you could read at 7am, and a report that says which words paid.

  • Once

    The account read

    Leaking terms, wasted match types, feed errors, tracking gaps, the fix order.

  • Week one

    The intent map

    Your terms in three tiers with the ad and the page for each. The document the account is built from.

  • Week one

    The feed, corrected

    Titles, attributes and images rewritten for how people search. Disapprovals cleared.

  • Every week

    The negatives file

    Every word we stopped paying for, and why.

  • Every week

    The four-minute report

    Spend, results, cost per result by tier, what moved and why.

Receipts

A lead-gen account, €38.14 to €2.47 a lead.

  • €38.14 → €2.47

    Cost per lead on a health-insurance account in France. 26,389 leads in ten months, revenue €4M to €8M in a year.

    Real client account · identity withheld · Meta Ads account, Sept 2026

  • 750+

    Funnels built between us. We paid for the failures ourselves.

    Founders’ own history · self-reported

  • $45M

    Revenue across the two of us, $35M in e-commerce and $10M in lead generation, with $10M of Meta spend behind it.

    Founders’ own businesses · self-reported

A real account

€38.14 a lead when they arrived. €2.47 now.

Real client account · identity withheld · read from the Meta Ads account on 20 September 2026 · national campaign, platform-attributed website leads · figures in EUR

  1. Before

    A health-insurance lead-generation account in France. Every lead cost €38.14, for about 57 leads a month. Five people on the sales team, and the cost per lead decided whether the month worked. The founder was the one checking Ads Manager at night.

  2. What we changed

    The same three stages every account gets. We read the account, the landing page and the best ads, and wrote down where the money leaked. Then tracking was fixed, the structure rebuilt, the first creatives shipped. Since then, every week: new creative in, losers out, budget to the winners, with a note on what moved and why.

  3. After

    A lead costs €2.47 today, about 1,028 a month, and the record month came in at €2.46. That is 15 times cheaper per lead, 26,389 leads in ten months, revenue from €4M to €8M in a year, and a sales team that grew from 5 to 15 people to keep up with the calls.

  • €38.14 per lead, when they arrived
  • €2.47 per lead, now
  • €2.46 per lead, record month
  • 15× cheaper per lead
  • 26,389 leads in ten months
  • 15 sales seats, up from 5

Your account is not this account, and we will not pretend it is. What carries over is the method. The session is where you find out whether it applies to you.

What the build looks like

Your ad, above the fold, on the words that buy.

Schematic, not a client account.

Thirty minutes with Jack or Max. Your account, read live.

Book a Profit Leak Strategy Session

This is for you if

  • A product people already search for by name or by problem
  • A store with a catalogue that Shopping can sell
  • A service business where a search means someone needs it this week

Not yet if

  • A brand-new category nobody searches for yet. Meta creates the demand; search catches it

Questions founders ask first

Before you decide.

Search or Shopping first?

For a store, Shopping first: the feed does the targeting. For a service, Search first, buyer-intent tier only, then widen.

Do you handle Performance Max?

Yes, as one campaign with its own budget and its own read, never as the whole account. It hides too much to run blind.

Minimum budget?

It depends on what a click costs in your auction. We tell you the number on the call after we have looked at your terms.

Can you run Meta and Google together?

That is the usual setup: Meta to create demand, Search to catch it. One report, both channels, one cost per result.

Next step

Thirty minutes. Your search terms, read live.

One of us opens your account with you and shows you which words are eating the budget.

Book a Profit Leak Strategy Session

30 minutes with Jack or Max. We read your account live and tell you where the money leaks.

Jack Thompson Maxime Le Morillon Jack Thompson & Maxime Le Morillon, founders

Prefer no call? The Crank Audit takes 90 seconds to apply and lands in 24 hours.

Not ready for a call?

The manual covers the rules we run every paid channel on: numbers first, creative second, tactics last. Free, 40 pages.

The Meta Ads Operating Manual, a dark hardcover with a brass turbine on the cover

The other lanes

Start with one. Most founders hand over the lot.