Give me a business and a number to hit. I find what is holding it back, whether that is the offer, the creative, the tracking or the operation behind them, then build or rebuild that piece until the number moves.
Compression wear · Egypt · roughly 40 days
Those are confirmed orders rather than delivered ones. A third of the orders placed were cancelled before confirmation, and Egyptian cash on delivery keeps refusing after that, which puts the delivered return closer to 11.5x.
Scroll the dashboard sideways to read it all.
Home water filtration · Egypt
Scroll the dashboard sideways to read it all.
Herbal wellness · Palestinian market · first month
This one ran inside the client's own ad account, so there is no dashboard of mine to show.
Three launches in, the question that decides the next one never changes: what is worth testing and scaling now. Most people answer it from instinct and a scroll through competitors' ads. I got tired of answering it that way, so I built something that answers it from measured demand across an entire supplier network rather than from one store's own sales.
It updates itself continuously behind a verification gate that refuses to publish when the data does not check out, and it returns no answer at all rather than extrapolating from thin history. That refusal was a deliberate design decision: a number nobody can stand behind is worse than a blank space.
It is running now, so rather than describe it, here is a login.
Understanding a business before spending money inside it is the same discipline whether the money is an ad budget or a software build. A system that does not match how the business really runs gets worked around within a month, so I research it first, map what happens today, find where value leaks out of the process, then specify the screens against that. Twelve industries researched, seven designed.
Two working prototypes
Each one shows the workflow analysis alongside the screens it produced, so you can see what the design was reasoning from rather than just what it looks like.
Healthcare
Admissions and approvals, bookings, clinic scheduling, the bed map, inpatients, critical units, medical orders and nursing procedures. A regulated operation where the paperwork is the process, and where a missing approval number is a bill nobody can collect.
Open the prototypeRetail · Saudi Arabia
Supply and suppliers, items, costing and distribution, payments, certificates, and clearance pricing. Built around seasonality and the cash problem that arrives with it, where last season's stock is this season's liquidity.
Open the prototypeThree things carry every launch above: how the campaigns are structured, the measurement that makes the numbers worth quoting, and the operation that has to survive the volume when the budget grows.
Campaign structure
Campaign naming is not housekeeping. Test and scale live in separate campaigns, the daily budget tier is written into the name, and each angle is isolated so it can win or lose on its own. Six months later it is still obvious what produced a result and what merely happened at the same time.
Scroll the table sideways to read it all.
Scroll the table sideways to read it all.
Measurement
Meta Pixel with the server-side Conversions API, GA4 with Tag Manager, and a UTM taxonomy consistent enough that revenue traces back to the campaign that produced it. None of the figures on this page would mean anything without that layer, which is why it gets built at the start rather than reconstructed at reporting time.
Operations
Order confirmation, courier handover and stock control, set up so that scaling the ads does not break what sits behind them. In a cash-on-delivery market this is where a good campaign quietly turns into an expensive one: orders that cannot be confirmed, stock that runs out mid-flight, couriers that return a third of the boxes.