From audit to scale

The truth nobody tells you in the first meeting: most accounts that come to us don't need a "better ad". They need someone to open the hood and see where the money leaks.

So our audit isn't a PDF with five notes. It's a forensic exam of the account, the site, the creative and the market, all in numbers.

First, the ad account itself. The last 6 months are read campaign by campaign, not as a summary. Every campaign gets a written verdict: proven-winner / message-cheap-deep / killed-bleeder. We surface what hides behind blended ROAS: campaigns optimized on photo-view instead of purchase (in an electronics case we found 62 of them, wasting a large share of spend), campaigns where IC exceeds ATC (tracking that makes no sense), frequency at 4 with creative fatigue, and every event set as Primary (we see it in 90% of Google accounts).

Second, tracking. Where is the Pixel, in the head or GTM only? Is CAPI live, is event dedup correct, is there one primary conversion or is everything primary? If the data is wrong, every decision after it is wrong. It's the first thing fixed.

Third, the destination. Where does the ad send people? Messenger or a Google Form means the funnel ends in a chat with no tracking (a developer case: 100% of spend to a form). A website gets a full CRO audit: speed, bounce, checkout, mandatory fields, shipping, reviews, CTA.

Fourth, creative. How many ads share copy? (13 of 33 in a skincare case, CPM up 10–25%.) Which angles have run for months and which burned out? Is there format variety or all static?

Fifth, the market. We pull your competitors' ads from the Ad Library, not a sample, and compare what they spend on versus what you spend on.

Finally a Health Score out of 100, and the leaks ranked biggest to smallest, in money.

You receive a written Findings file within 48 hours. Not an opinion: a ranked list of where your money leaks and what to fix first. If the audit shows we're not the answer, the file says so. It has happened before and will again.

The market tells you everything, if you know how to read it.

Our market research isn't a 40-page report nobody reads. It's four layers that produce a decision.

Layer one: competitors by their ads, not their claims. We pull every live ad of every competitor from the Ad Library. An ad running six months sells; an ad killed after a week burned. We extract angles, offers, prices, formats and recurring hooks. Real case: one video stayed a winner for six months with 1.8M views in a hyper-competitive niche, because it sat on an angle nobody else used.

Layer two: the audience in its own words. Competitor ad comments, product reviews, Facebook groups. That's where the real objection lives (shipping too expensive, afraid it won't look like the photos, no Arabic) before you spend a unit of budget.

Layer three: pricing and offers. Who sells at what, installments or not (in KSA a large share of orders run through Tamara; a retail case turned that into one of its highest-ROAS campaigns). The expected AOV, and the break-even ROAS you must clear to profit.

Layer four: a SWOT, but real. A strength is something you have, your competitor doesn't, and it shows in the ad. An opportunity is an angle the whole market left untouched. A threat is a competitor spending double on the same keywords.

Then one sentence: where you enter the market, with which message, at which price.

You receive a written market map, a competitor-ads sheet, and the angles we'll start with.

A plan without numbers is a wish.

The 90-day plan starts from three numbers of yours: margin, AOV and break-even ROAS. Once you know your AOV and margin, there's a ROAS below which you lose even with a green dashboard. That number goes on the plan's first line.

Then forecasting: to hit the target at a profitable ROAS, how much to spend, on which channel, in which week. Real case: an owner asked for a fixed sales target in the last 10 days. The offer was built on six months of data with a high AOV as the first condition. The plan set the number; the owner called to say "spend more"; the answer was no; the target was exceeded.

The funnel is split into stages, each with its own budget and KPI: TOF (new, cost per engaged user not sales), MOF (warm, ATC and CPC), BOF (added to cart and didn't buy, CPP), retention (bought, repeat rate).

Creative is requested in the formats the data points to, not the ones the designer likes. Seasons go on the calendar two months early, not a week (a Black Friday case hit 60% of target in the first 10 days because the offer was prepared two months earlier).

You receive a written 90-day plan, a media plan with a weekly forecast, and a creative calendar. And you can hold us to it every Sunday.

Every platform tells you it brought the sales. Meta says so, Google says so, TikTok says so. And their sum is far above your real revenue.

That isn't cheating. That's attribution. And until someone fixes it, you keep raising the budget of the wrong platform.

What we do before any launch: Pixel in the head, not GTM only, so it fires on first paint; CAPI with correct event dedup so nothing counts twice; a single primary conversion (Purchase), not every event primary; UTMs per platform and campaign; and MER (marketing efficiency ratio) as the governing number, total revenue over total spend, not one platform's ROAS.

Real case: one week and 9,008 orders. The first problem solved was every platform claiming the sales; after UTMs and Mixpanel, TikTok and Google emerged as the strongest, and budget was allocated accordingly.

After tracking, the dashboard: one board gathering every channel, the CRM and the store. Open it on your phone and see spend, return, MER, and which campaign is bleeding right now.

And a layer called market intelligence that never stops: competitors' new ads, angles starting to repeat, price changes. It arrives in the weekly report.

You receive verified tracking (event screenshots), a live dashboard, and a weekly intelligence report.

This is the daily work, the part where agencies disappear.

Structure: not every campaign in one bucket with the same goal. An evergreen CBO bringing new people without an offer; a testing ABO with every offer or angle in its own ad set; a scaling CBO for winners only; DPA split into product sets (category, price, best sellers); retargeting split into visitors / ATC and IC / 90-day buyers excluding the last 30 days.

Creative: from one message, five angles, each a hook. A test every 7 to 10 days with written directions for the designer. Worked? A static on the same angle and a second test. Failed? Metrics read before discarding. Whitelisted ads from the creator's own account cut CPA 30% with the same video in a real case.

Scaling rules, written: flat daily budgets instead of jumps (a real-estate account on a flat daily budget held CPL low; the jump-driven test cost eight times more per lead). +20–25% every 3–4 days once stable; 50–100% on paydays and seasons with Target ROAS. No punishing a campaign at day's end; dawn brings 20–25% of daily sales.

Bleeders killed before they grow, by a written rule: a campaign below break-even for X days at Y spend gets killed. Not a mood decision.

Daily follow-up: every morning, yesterday's spend, MER, campaigns past the threshold, and the action. Every Sunday at 10am, four numbers on your WhatsApp: spent, returned, killed, next.

You receive an account audited every day, and a Sunday report you understand in one minute.

"I scaled hard today, we got 1,000 orders, and CPA is way down!" ... "Good. Pause the ads now."

That exchange is from a media buyers' group and repeats every week. A thousand orders in a day on unprepared operations means shipping delays, cancellations and returns. The ROAS on screen isn't the cash that reaches the bank.

So our scaling is tied to capacity, not CPA alone. Before raising, we ask: how many orders can stock carry, how many days does shipping take, who answers the messages? Real case: a fashion brand grew sales many times over in three months while ads were paused four times because stock ran out; triple the number was possible with fulfilment ready.

How we actually scale. Horizontal: new channels once the first is stable. TikTok in the Gulf is a gold mine (cost per booking 16 on TikTok vs 60 on Meta in a tourism case); Google for existing demand (ROAS 41 in one quarter); Snapchat for Gen Z in the Gulf; WhatsApp as a scheduled sales channel (three times a month in a food case). Vertical: higher AOV with bundles and shipping tiers (AOV-based shipping tiers cut abandonment and held AOV in a record-sales case). Deep: retention, because a past buyer is the cheapest customer you'll sell to again (+50% in the slowest month of the year with no offer, loyalty only).

Seasons are prepared two months ahead on the calendar.

You receive capacity-based scaling rules, a channel plan in order, and growth that continues after the good month.

The goal isn't a good month. The goal is your competitor starting to copy your ads; that's when you've arrived.

At this stage the work shifts from campaigns to a system you own.

Systems and automation: a CRM that scores leads on its own, reports written from data every day, WhatsApp flows that bring back cart abandoners, and a catalog that refreshes itself (a campaign ran for months without dropping because the bestseller tag updated daily from Shopify Flow).

AI on your company's data: an assistant replying to customers 24/7 in their dialect, scoring leads hot or cold, drafting copy and creative from winning angles, and analyzing campaigns to flag a problem before it becomes one.

Your team: we train your team in the GP MAX academy on the same system we run. 500+ graduates who now run campaigns themselves. Stop renting marketing; own it.

The moat: content and social proof that grow awareness and ease the buying decision (in a Google case, what mattered most was the social funnel, branch videos and reviews, not bidding); retention that makes weak months not weak; and new markets opened by rule: three local creatives, a capped test budget, a warm audience from day one (the rule born from a failed Gulf-test).

You receive a growth system you own, even if we part ways.