Meta Ads for webshops
Facebook and Instagram for DTC brands, run as two separate jobs: finding people who don’t know you yet, and being worth coming back to. Most accounts blur the two and then wonder why the numbers move sideways.
Scope
- Prospecting
- core
- Retention / retargeting
- core
- Creative testing
- structured
- Catalogue / Advantage+
- with guardrails
- CAPI & event quality
- included
One-off audit, creative test sprint, or monthly management. Audit first if we haven’t worked together.
How I work — in order
-
Fix the signal
Pixel and Conversions API first: if Meta is guessing at your events, everything after it is guesswork too.
-
Separate the jobs
Cold and warm in their own campaigns, with budgets you can defend separately.
-
Test creative, not settings
One variable per test — hook, format or offer. Settings rarely move the needle; the ad does.
-
Report what happened
Which creative won, by how much, and what it means for the next batch.
What you get
- A clean split between prospecting and retention spend
- A creative test log: what ran, what won, what it cost to learn
- Event quality high enough to trust the reporting
- Full ownership: your Business Manager, your pixel, your data
What I need from you
- Admin access to the Business Manager and ad account
- Whatever creative you have — stills, video, UGC, product shots
- Your real margins, so we can judge a winner properly
- Enough budget per test to reach a conclusion, not a hunch
Fair questions
About Meta Ads specifically.
Do you make the creative too?
I don’t shoot or design. I write the briefs, decide what gets tested against what, and tell you which of your existing assets to cut. If you need production, I’ll point you at people I trust.
Advantage+ or manual campaigns?
Both, depending on catalogue size and how much signal you have. What I won’t do is hand everything to automation and call the resulting ROAS a result — the same brand-traffic problem as Pmax applies here.
My ROAS dropped after iOS changes. Can that be fixed?
Partly. Server-side events and better deduplication recover a real share of it. The rest is a measurement problem, not a performance problem, and it’s solved by comparing against your back office rather than chasing the number in the dashboard.
Spending on Meta without knowing what the creative did?
Mail me the account and the last three months of results — I’ll tell you what’s worth testing first.