home / work / current client work

current and recent clients

These come from work I'm doing now or did recently, so client names and sensitive numbers stay private.

current client worksubscription appsgamescommerce

recent clientengagements.

The projects I'm working on now, or finished recently. Each one covers what the business needed, what we found, what we changed and where it stands. Some are still in progress.

MetaApple Search AdsGoogleweb-to-appmeasurement

AI coaching subscription app · iOS, Android and web · six-figure monthly paid spend · 2026

Building an in-house paid engine for a fast-growing AI app

the brief

Bring paid media in-house from an agency, and build a system the team could own: campaigns, creative testing, measurement and reporting.

where it stands

Purchase-optimised web campaigns showed a reported return above 1x early on. The bigger result is a testing and reporting system the team can keep running.

what we found

Copying the agency's best campaigns would have put two sets of the same ads into the same auctions.
Creative tests weren't getting fair spend under automatic budget allocation.
A drop in web trial-to-paid turned out to be new campaigns still learning. The same funnel on another channel hadn't moved.
A design update had buried the iOS tracking permission prompt, quietly hurting new app campaigns.

what we changed

A separate in-house learning lane, using the agency's results as a baseline instead of a template.
Fixed budgets per test, so every creative got a fair read.
Separate optimisation lanes: trials, qualified trials, direct purchase, and purchase with a minimum return.
Weekly reporting split by app, web and total, plus a predicted net-ROAS model with clear rules on when it can change.
New creative formats, including chat-style and founder-led ads.

MetaApple Search AdsGoogleASOSEO

Subscription app for small business owners · since late 2025

Making paid acquisition pay off for a business app

the brief

Scale paid acquisition profitably across Meta and Apple Search Ads, and get tracking and store pages to a point where the numbers could be trusted.

where it stands

Install-to-trial improved, and spend no longer leans on a single ad. Ongoing.

what we found

Meta's cost per trial was rising while ad costs were falling, so the problem sat after the click, mostly in one placement.
Cheap test installs were producing almost no trials.
About half of search ad installs came from people searching the app's own name.
Apple and the attribution tool disagreed on installs, partly because tracking only started after an in-app privacy prompt.
Some ads promised one thing while the App Store page said another.

what we changed

Removed the weak placement from the trial campaign.
A three-lane creative system: install screening, then trial challengers, then core scaling.
Brand and non-brand search judged separately, and Apple's broad keyword suggestions filtered by hand.
Custom product pages matched to each ad's promise.
Tracking fixes, Android trial events, and later SEO and AI-search visibility work.

Metafunnelmeasurement

Desktop software for creative professionals · web sign-up · 30-day free trial · 2026

Deciding whether paid ads should run at all

the brief

Work out whether Meta was worth the money, and how to scale it if it was.

where it stands

Campaigns are paused while budget is limited. A 14-day trial has been tested; my recommendation is shorter still. Spend restarts once the funnel and data are ready.

what we found

Sign-ups looked healthy, but trial-optimised campaigns had brought almost no paying customers.
With a 30-day trial, payment lands weeks after the click. Meta's 7-day click window almost never sees it, so it can't learn who pays.
Too little purchase data was flowing back to Meta to optimise on.

what we changed

Recommended fixing the funnel and the data before spending more.
Recommended testing a 3-day or 7-day trial, so payment lands inside the window Meta learns from.
Pointed optimisation toward early activation, which predicts paying.

Metameasurementcreative

Fantasy sports game · web and iOS · through an agency partner · 2026

Launching paid acquisition for a fantasy sports game

the brief

Launch paid acquisition for the NFL season, set up measurement, and find out which channel brings real players.

where it stands

Early days. Payments are the next event to optimise toward, once they're reliable.

what we found

Web sign-ups were cheap, but those players rarely came back.
App installs were even cheaper and retained almost like organic players.
Licensing rules meant no league logos or player images in ads.

what we changed

Started web-first because measurement could be ready fast, with the Pixel and Conversions API in place.
Moved the weight to app campaigns once the retention data came in.
Judged channels on a ladder of events: sign-up, install, first game, repeat play, payment.
Creative built around real gameplay on a phone screen, with players reacting on camera.

creative strategyMeta

Travel, fitness and AI fitness apps · through an agency partner · 2026

Creative strategy for three subscription apps

the brief

Write the strategy, scripts and briefs behind weekly paid social creative, working with designers, creators and the client.

where it stands

Ongoing weekly creative across all three apps.

what we found

RV travellers weren't buying 'camping'. They were careful planners paying to avoid low bridges, bad stops and wrong listings.
Price-heavy scripts were drowning out what the product actually does.
Long, complicated briefs slowed down a one-designer team.
A more personalised web quiz didn't beat the simple app path.

what we changed

One worry per ad, then the exact app screen that solves it.
One simple offer per campaign, with creative built so prices could be swapped quickly.
Shorter scripts and clear briefs, using only footage that actually exists.
Variations on proven winners instead of endless new concepts.

TikTok Shopunit economics

Seed and plant store · TikTok Shop ads · 2026

Testing TikTok Shop for a seed and plant store

the brief

Run TikTok Shop ads and find products that could scale profitably.

where it stands

Paused. More creative couldn't fix product mix, margin and season, and saying so early saved the client money.

what we found

One product carried most of the account, and only about five had been tested.
The same product was profitable as a product card and unprofitable through one affiliate video.
Return fell below break-even even with fresh creative, as the season turned.

what we changed

Judged every product on its own numbers instead of the account average.
Proposed widening to 12 to 15 products with several creatives each, and raising return targets in steps.
Paused campaigns when return dropped below 1.

want this on your app?

let's find what's holding your growth back

A Growth Diagnostic looks at your ads, funnel, pricing and tracking together, and tells you what to fix first. Five working days.

See how it works Case studies

Current and recent client work, shared without names or confidential figures. Some projects came through an agency partner.