Summary
I reviewed 90 days of your Google Ads data. The account is actively managed but is working well below its capability. The problems are in how it's built, not how hard it's worked.
Six areas of work, sequenced over 90 days.
1. What the audit found
Five findings. Detail and figures are in the full audit; every number traces to your own exports.
1.1. Budget is buying demand you already own. $18,617 over the 90 days, 85.9% of your Performance Max budget, went to people searching for you by name. It is bought at open-auction prices through a campaign not labelled as brand, with no control over the message shown.
1.2. The account cannot distinguish a large opportunity from a small one. The dominant conversion carries no value, and every campaign bids to cost-per-conversion. The system is instructed to find the cheapest lead. Cheapest correlates with smallest.
1.3. Campaign types are being compared unfairly. Purchases are only counted in your Performance Max campaign. Your search campaigns drive purchases too, but aren't set up to record them, so they look worse than they are. Every budget decision made from that comparison rests on a false picture.
1.4. Two categories you publicly sell receive no paid budget: YouTube channels and digital agencies.
How this was observed: I listed every enabled ad group and keyword in the account, cross-referenced that against the categories you market publicly, then searched the 90-day search-terms report to check whether demand was arriving regardless. Finally I traced where those queries were being matched and where the ads pointed.
- YouTube channels. No seller-side ad group or keyword exists anywhere in the account. The only YouTube ad group sits inside the US - Buyer campaign, at $276 across the 90 days. Seller queries are being matched into it (
sell my channel,selling youtube channel,yt channel sell,youtube sell) and served buy-side ads pointing at a category page shared with the Instagram ad group. Someone trying to sell a channel is being sent to the buy side of your own marketplace. - Digital agencies. Zero keywords and zero ad groups, account-wide. The demand still arrives, but only as loose-match leakage into unrelated ad groups:
digital marketing agency for salematched into a Blog ad group,web design agency for saleinto a generic Business ad group, andselling a home care agencyinto Valuation - SaaS.
You are already paying for this demand. It's simply being matched into places with no ad and no page written for it.
1.5. Traffic and landing pages are misaligned. Specific, expensive intent lands on generic pages. Three UK ad groups send traffic to your homepage rather than the valuation tool their US and Australian equivalents use: $2,212 over the 90 days, 30% of the UK budget. Each of those three ad groups also holds paused ads pointing at a dedicated seller landing page, so the better destination already exists and isn't being used.
2. The six work areas
| Area | Outcome | |
|---|---|---|
| 1 | Measurement | Bidding pursues value, not volume |
| 2 | Demand capture | Brand demand captured deliberately, not at auction prices |
| 3 | Category coverage | Every category you sell gets paid coverage |
| 4 | Landing pages | Page matches query, proven by testing |
| 5 | OpenAI (ChatGPT) Ads | Early position in a channel priced below your Google terms |
| 6 | Account management | Checked, documented, auditable |
Area 1: Measurement
Problem. The account optimises toward the cheapest lead available. Cheap leads skew small, so the bidding works against a business whose margin sits in larger deals.
Work.
- Attach commercial value to the lead so bidding targets larger opportunities
- Correct conversion counting so numbers describe people, not repeat sessions
- Connect the ad platform to what eventually qualifies and closes
- Migrate primary measurement onto native tracking
Why you're well placed. Your valuation tool already asks prospects what their business is worth. That signal exists and isn't being used.
Expected effect on reporting. Conversion counts will fall and cost-per-conversion will rise when this is done correctly. The numbers will look worse because they'll finally be right. It needs stating in advance.
Area 2: Demand capture
Problem. Brand demand is currently bought at auction prices, without message control.
Distinction that matters. Paying to intercept someone who would have found you anyway can be seen as waste. But any investment in awareness (PR, social, content, partnerships) surfaces later as someone searching your name. That search is where the awareness spend is collected. Uncontrolled, it is either bought back expensively or intercepted by a competitor.
Who is actually searching your name. A brand search is not the same thing as an existing customer. Many of the people typing your name have never transacted with you. They heard about you from a press mention, a podcast, a forum thread, an AI assistant, or someone they know who sold a business, and they are using your name as a shortcut to the category rather than returning to a service they already use. Treating all brand traffic as demand you already hold is the assumption worth testing. If a meaningful share are first-time prospects who arrived through awareness activity, that traffic is prospecting, not repeat business, and it should be costed accordingly.
What removing brand entirely does. Worth knowing before the decision is made permanent:
- Reported conversions fall by roughly 21%, and Performance Max moves from about $40 per conversion to considerably worse. That is the brand subsidy becoming visible, not performance changing.
- Actual demand lost is far smaller than that figure suggests, because your organic listing still catches most people searching your name.
- What you give up is control and defence. No ad copy written for a high-value seller, no sitelinks routing them toward the right product, and no cover if a competitor decides to bid on your name.
None of that argues for switching brand back on today. It argues for keeping the option open until the account can show you who those searchers actually are and what they are worth.
Work.
- Stop the uncontrolled purchase first
- Establish the true cost of the non-brand engine once that demand is separated out
- Decide deliberately, once the account can see deal size, whether a capped brand presence earns its place
- Check whether brand traffic is being routed to messaging built for the deal size you want
Area 3: Category coverage
Problem. Each category has its own buyers, its own search language and its own competitors. Some are funded properly. Two get nothing, despite demand actively arriving.
Work.
- Build coverage for every category you sell, on the correct side of the marketplace
- Base it on demand already visible in your own search data
- Structure it as a repeatable template so each new category is a rollout, not a project
Area 4: Landing pages
Problem. Traffic bought on specific intent lands on generic pages. Your account already shows what this costs: your buyer campaign is the only one with pages matched to the query, and it converts at $10 against $37 to $64 everywhere else.
Work.
- Correct traffic landing where it shouldn't
- Build genuine page depth using infrastructure already in place
- Test properly
Testing, realistically. At your traffic levels, large differences (a different page, a different proposition) read within a fortnight. Small tweaks cannot be read inside a quarter. So: test propositions, not components, and pool traffic so results arrive in weeks rather than months.
Highest-value test available. Not a design question but a positioning one: whether broker-led framing converts better-quality sellers than marketplace-led framing. A strategic question settled with money instead of opinion, in roughly two weeks.
Area 5: OpenAI (ChatGPT) Ads
Problem. Contested Google keywords cost what they cost, and no amount of account management makes them cheaper.
Meanwhile the research phase that precedes a business sale has been moving into AI assistants: weeks of "should I sell, what's it worth, who should I use, is this platform credible". That traffic didn't disappear. It relocated.
The commercial case, in one line. OpenAI's recommended cost-per-click bid is currently $3 to $5. That price reflects how few advertisers are bidding, and it will rise as more arrive. Here is what you already pay on Google, search campaigns only and non-brand:
| Avg. CPC | |
|---|---|
| US - Seller | $4.01 |
| US - Seller - Combined | $3.72 |
| Your US seller-side blended CPC | $3.90 |
sell shopify store | $8.62 |
[sell my website] | $8.59 |
sell business | $6.14 |
Every brand keyword in those campaigns is paused, and the brand fragments that still slip through close variants account for 0.33% of their spend.
You are already paying the price of entry. Being early to this channel carries no premium.
That is the entire argument. It's the same click, earlier, before the auction gets bid up.
Limits I'd state up front.
- Volume is small. Ad load is deliberately light, so this cannot absorb serious budget. It is a position, not a destination.
- The attribution window is 30 days, against a sales cycle measured in months. It can measure a lead well; it will never measure a closed deal.
- No published benchmarks exist, so we measure against your own Google performance rather than an industry figure.
Why me on this specifically. I've been running and testing this platform hands-on rather than reading about it, so I know what earns its spend and what doesn't. With no published benchmarks to work from, most advertisers entering it are guessing.
Area 6: Account management
Problem. Defects reaching the client before the agency.
Work.
- Scheduled checks that ads point where they should, in every market
- A real pruning cadence rather than accumulated waste
- Monitoring for campaigns quietly capped or policy-restricted
- A readable change log
- Reporting that answers "is this working", not "what did you do"
3. How performance would be measured
| Measure | Why |
|---|---|
| Cost per genuine lead, counted honestly | Replaces a figure inflated by repeat sessions |
| Non-brand cost and volume | First clean read of the engine |
| Proportion of leads at the deal size you want | Measures the commercial goal directly |
| Cost per qualified opportunity | Replaces a target set against a free tool |
| Category coverage | Closes the gap in 1.4 |
4. Background
A note on where the judgements above come from, since several of them are calls rather than calculations.
The practice. Google Ads specialist since 2015 and a certified Google Partner since February 2015. Around $15M in managed ad spend, 100+ clients, 15+ industries. Search, Shopping and Performance Max certified. No agency layers and no junior staff, which means you would be dealing with the person who ran this analysis rather than an account manager relaying it.
Where the recommendations come from. None of the problems in this document are unusual, and none of them are new to me. The work that recurs across the accounts I take on: restructuring so that bidding can see value it was previously blind to. Finding that most of a budget was flowing through structurally broken campaigns while the account's own best performers sat blocked. Rebuilding lead generation funnels where the landing experience, rather than the traffic, was the constraint. Running multi-market accounts where each market needed its own targets instead of one shared number. Every one of those has a counterpart in Section 1.
On ChatGPT Ads, I have been running it rather than reading about it. That is the only reason I would put a new channel into a document like this one.
5. What's Next
Regardless of what you decide. Two of the findings in Section 1 can be checked by your current team this week. One of them, traffic landing on the wrong page in one market, is a twenty-minute fix that recovers spend immediately. You don't need me to act on it.
If we proceed, the first two weeks need no additional budget. They recover money already being spent badly and give you a baseline you can trust. Nothing gets scaled until the measurement is trustworthy.
| Weeks | Focus | Budget required |
|---|---|---|
| 1 to 2 | Fix what's visibly broken | None; recovers existing spend |
| 2 to 6 | Make the commercial goal measurable | None, but requires some engineering support |
| 4 to 10 | Fund unfunded categories; build page depth | Reallocation |
| 8 to 12 | New markets and the channel test | Incremental, capped |
What I'd need from you.
| Why | |
|---|---|
| Read-only account access | Impression share, auction data, campaign settings and asset reporting aren't visible in exports. Everything above was reconstructed without them; it would be firmer with them |
| A small amount of engineering time | For Area 1, the highest-return work in the plan, and roughly a day of it |
| Your lead, qualified and closed rates, by category and market | Turns a proxy for value into an actual value |
| A route to deploy landing pages | The infrastructure exists; the gap is page depth |
| Clear ownership of the account | If more than one party is working in it, hard-fenced campaigns and separate budgets, in writing |
INVESTMENT
$5,000 USD/month
Covers all six work areas above. Billed monthly. No lock-in beyond the initial 90-day engagement.
Decision point at month three, taken on the leading indicators in Section 3, not on closed revenue, which a sales cycle of your length cannot produce inside the engagement.
Based on 90 days of Google Ads data, 24 April to 22 July 2026. Full findings and figures available in the detailed audit. No changes have been made to the account.