Paid ads management
We run your Meta, Google and TikTok campaigns end to end — the buying, the creative, the testing and the scaling — and we judge all of it on cost per purchase rather than on clicks.
What breaks
Spend concentrates where measurement still works
When conversion data is patchy, the platforms optimise toward the events they can still see. That is usually the cheap, top-of-funnel action rather than the purchase, so cost per click looks healthy while cost per purchase quietly climbs. Most accounts we open have been drifting this way for months without anyone noticing, because the dashboard metric everyone watches is the one that still looks fine.
Creative fatigues before anyone reacts
Performance decay is gradual, so it reads as a bad week rather than a dead asset. By the time frequency is obviously too high, the account has usually spent weeks paying a premium to show a tired ad to an audience that has already dismissed it. The fix is almost never a bigger budget on the same creative.
Structure fights the algorithm
Accounts accumulate campaigns the way drawers accumulate cables. Overlapping audiences bid against each other, budgets fragment across too many ad sets to ever exit the learning phase, and old campaigns keep spending because nobody remembers what they were for.
Testing has no controls
Changing creative, audience and budget in the same week produces a result nobody can attribute. Without a stable baseline and one variable at a time, an account generates opinions rather than evidence, and the loudest opinion wins.
- Fragmented
- Twelve ad sets splitting one budget, with audiences overlapping so campaigns bid against each other. None receives enough spend to exit the learning phase, so the platform never gets a stable signal to optimise against.
- Consolidated
- Three well-funded campaigns with deduplicated audiences, drawing from one creative pool. Each clears the learning phase and stays there, so optimisation compounds instead of restarting every week.
Illustrative of the structure, not a screenshot of any account.
How we find it
Account and structure audit
We map every live campaign, ad set and creative against what it actually spends and returns, then identify overlap, redundancy and anything still running that no one can justify.
Measurement check first
Before we touch a bid, we confirm the events firing, their deduplication, and whether reported conversions reconcile with orders in the store. Optimising against untrustworthy numbers just makes the wrong thing happen faster.
Creative and fatigue analysis
We look at frequency, decay curves and hook retention to separate assets that never worked from assets that worked and have since worn out. They need opposite responses.
Baseline agreed in writing
We record the starting position and both sides sign off on it. This is also the number the performance guarantee is measured against, so it cannot be revised afterwards.
How we fix it
Rebuild the account around consolidation
Fewer, better-funded campaigns that exit learning and stay there. Audiences deduplicated so they stop bidding against each other, and everything that cannot explain its own existence switched off.
Repair the signal
Server-side events with proper deduplication and identity matching, so the platforms optimise against complete data. This is usually the single highest-leverage change, and it is invisible in every front-end report.
Ship human-made creative on a rotation
Concepts written for your product and your buyer, produced in-house, and replaced on a schedule set by decay data rather than by whoever remembers. Volume is planned in advance so there is always a fresh asset ready before the current one dies.
Test one variable at a time
Structured tests against the agreed baseline, with a decision rule set before the test starts. Results are recorded, so the account accumulates evidence rather than folklore.
Scale deliberately
Budget increases are staged so campaigns are not thrown back into learning. When something works, we widen it carefully rather than doubling it and hoping.
The operating rhythm
What actually happens, and how often. Published because “ongoing optimisation” is what everyone says and almost nobody defines.
- Daily
- Spend and cost-per-purchase check across every live campaign, with intervention only where the data warrants it. Most days the right action is none, and resisting the urge to tinker is part of the job.
- Weekly
- Creative performance review against decay curves, rotation of anything fatiguing, and a written update covering what changed, what it cost and what we expect from it.
- Fortnightly
- New concepts into production, briefed from what the current winners have in common rather than from a content calendar.
- Monthly
- Full account review against the agreed baseline, structural changes if the data supports them, and a candid read on whether the current spend level is still the right one.
Scope
What is included
- Full account audit and written baseline, agreed before any spend changes
- Campaign build, restructure and daily management across Meta, Google and TikTok
- Human-made ad creative — concepts, production and a planned rotation schedule
- Structured testing with the decision rule set before each test runs
- Weekly written update and a monthly account review
- Direct access to the person running the account, not an account manager
What we need from you
- Delegated access to your ad accounts through each platform’s permission system — never your password
- Advertising budget funded directly with the platform on your own payment method
- Product, brand and any existing creative assets you want us to work from
- A named decision-maker who can approve creative within a few days rather than a few weeks
- Honest numbers on margin, so we optimise toward profit rather than revenue
What changes after
- Spend is attributable — you can see which campaigns produce orders and which produce activity.
- Cost per purchase becomes the number the account is steered by, not cost per click.
- Creative is replaced before it decays, so performance stops sawtoothing.
- Budget changes are decisions with a reason attached, made weekly rather than reactively.
These describe the mechanism of improvement, not a guaranteed outcome. Results depend on your product, margin and market — the written guarantees are in section 7 of the Terms.
Questions
Do you need admin access to my ad accounts?
No. We work under delegated access granted through each platform's own permission system, at the minimum level the work requires. You keep ownership of the accounts, the pixels and the audiences throughout, and you can revoke our access at any moment without needing us to action it.
Who pays for the ad spend?
You do, directly to the platform. We never receive, hold or disburse your advertising budget — it stays on your own payment method throughout. Our fee is separate and invoiced independently, which also means there is no incentive for us to recommend spending more than the account can justify.
What if my current agency built the account?
That is common and it is not a problem. We audit what exists before changing anything, and where the previous structure is sound we keep it. Rebuilds happen where the structure is actively costing you money, not as a matter of routine.
How many campaigns should I actually be running?
Fewer than you are, almost certainly. The common failure is a dozen ad sets each receiving too little budget to ever leave the learning phase, several of them bidding against each other for the same people. Consolidation feels like losing control and is usually the single change that moves cost per purchase most. The right number depends on budget and catalogue, and the audit gives you a specific answer rather than a rule of thumb.
Will you pause campaigns that are currently making money?
Not without telling you first and explaining why. Where something profitable is also cannibalising another campaign or preventing consolidation, we will show you the overlap and let you decide. Nothing that spends money gets switched off silently.
Go deeper on the mechanism
Find out whether this is your bottleneck
The free diagnostic scans your storefront and tells you where the problem actually is — which is often not where it feels like it is.