Performance Marketing

Where is your money leaking? We name the exact stage.

Most accounts we review are not held back by the advertising itself. The problem is one point in the funnel quietly eating the spend: tracking reporting the wrong numbers, a landing page contradicting the ad, or a checkout losing people at the final step. We walk all twelve points against your own numbers, name the stage you are standing on, and give you the first three things to fix. Then we run the campaigns on that basis: Testing, Killing and Scaling, week by week.

See the twelve points

A full funnel report within five working days
Delivered whether or not we work together
Credited against your first month if we do

Your funnelillustrative

Visits10,000

Add to Cart3,200

Checkout890

leaking here

Orders740

72% drop between Add to Cart and Checkout. The problem is not the ad, it is shipping appearing late.

Results

Figures from client accounts, exactly as Meta calculates them.

The screenshots below come straight out of Meta Ads Manager. We have left Meta's own labels untouched, and under each figure we say what it measures, so you can compare like with like rather than by impression.

Account names are withheld under our agreements with these clients, and the detail rows are obscured so we do not expose their campaign structure. The totals and dates are untouched.

Ecommerce

Objective: orders · Meta Ads Manager

Purchase revenue in 19 days: 0EGP 1,532,861

Average ad ROAS (as Meta reports it): -40.61
Period: -1 to 19 June 2024
Business age at the time: -3 months

What happened

The business was only three months old when this ran, and the window is 19 days, not a full month. Note that the ROAS in the screenshot is labelled Average, meaning the mean of the individual ads rather than total revenue over total spend. We flag it so you can compare our number against anyone else's on the same basis, because the two are different figures and each reads differently.

1 to 19 June 2024: total purchase revenue EGP 1,532,861.04 · average ROAS 40.61

Source: Meta Ads Manager

Ecommerce

Objective: orders · Meta Ads Manager

Purchase revenue in the period: 0EGP 3,466,158

Average ad ROAS (as Meta reports it): -96.23
Lowest cost per purchase in the account: -EGP 1.77

What happened

3.4 million in revenue, with the best ads in the account buying a purchase for EGP 1.77. That figure happens when three things meet: a healthy margin, an audience tuned by testing rather than guesswork, and creative that stops the scroll. It is not a promise for every account, but it is the ceiling we can reach when the numbers allow it.

Total purchase revenue EGP 3,466,158.63 · average ROAS 96.23 · lowest cost per purchase EGP 1.77

Source: Meta Ads Manager

These figures come from ecommerce accounts. We have results on other objectives (leads, messages, bookings) and in other categories.

Want to know the expected cost range in your own category? The review works it out with your numbers.

Request the audit

The audit

We tell you exactly where the funnel is leaking

"The ads are not working" is never one problem. It is twelve different places money can leak, each with a completely different fix. The audit starts with tracking, because if measurement is wrong every number after it is wrong, then works through the ad account, then through the site itself as a full CRO pass.

Tracking

First, because every number after it is built on this

Events SetupAre events firing correctly? Purchase, AddToCart, InitiateCheckout?
The reading that points at itPlatform numbers disagree with your actual sales. If the gap is wide, you are optimising against data that does not exist.
Deduplication & Signal LossIs the event counting twice? Or are conversions disappearing entirely?
The reading that points at itPlatform conversions run higher than your real orders, meaning one event is counted more than once. Or they dropped suddenly after an update, meaning events are lost before they reach the platform, which happens when the Conversions API is not set up.
AttributionAre the windows aligned, and are platforms claiming the same sale?
The reading that points at itAdd up conversions across every platform and the total exceeds your real orders. Somewhere you are paying twice for the same sale.

Before the click

The part the ad account controls

CreativeDoes the creative stop the scroll, or not?
The reading that points at itHigh reach, low CTR. People are seeing the ad and not clicking, so the problem is the message or the format, not distribution.
TargetingIs it reaching people who could actually buy?
The reading that points at itHealthy CTR, visits that never convert. The ad pulled people in, but people who browse rather than buy.
PlacementIs it showing where results actually happen?
The reading that points at itPerformance varies sharply between placements, and part of the spend is going to placements that never produced a conversion.

After the click: CRO

The part the site controls, and where most of the leaking happens

Landing or product pageDoes the page keep the promise the ad made?
The reading that points at itCheap clicks and fast exits. Time on page is very short, usually because the ad promised one thing and the page says another.
Page Speed & MobileDoes the page open quickly on an ordinary phone and connection?
The reading that points at itExit rate is far higher on mobile than on desktop. Most of your spend goes to mobile, which multiplies the damage.
Visit to Add to CartIs the product itself convincing? Price, images, description, trust?
The reading that points at itPeople reach the page and never add. The problem is the offer: the price, the images, missing information, or no reason to trust you.
Add to Cart to CheckoutIs something surprising them in the cart?
The reading that points at itThey add and abandon. Most often shipping appears late, or the final total came out higher than expected.
Checkout to completed orderIs paying easy, and is there a reason to trust you first time?
The reading that points at itThey reach payment and stop. Too many steps, forced registration, a missing payment method, or no reviews and no clear returns policy.
AOVCan the order value carry a customer acquisition cost at all?
The reading that points at itEverything above is working and the numbers still do not close. The problem is not the advertising, it is that average order value is too low.

This is why the audit is a paid piece of work. It is not an automated tool report. We go into the ad account, the tracking setup and the site data, and walk all twelve points against your own numbers. You get a report naming the stage that is leaking, the first three things to fix, and the order to fix them in, ranked by impact rather than by effort.

Objectives

What we optimise toward

Buying an order is not buying a lead is not buying a conversation. Each objective is a different campaign structure, a different optimisation signal, and a different definition of success. First we establish what you are actually buying.

Orders

The metricROAS, cost per order, average order value

Who it suits: Online stores and brands selling direct
What changes in the work: We connect the catalogue and configure purchase events, and optimisation runs on order value rather than order count. A campaign delivering 100 orders at an average of 200 is weaker than one delivering 60 at 700.

Leads

The metricCost per qualified lead, not per lead

Who it suits: Real estate, education, professional services, B2B
What changes in the work: Filtering happens before the form, not after it. We add qualifying questions and exclude audiences that will not buy, even when that raises cost per lead, because a cheap lead that never buys gets paid for twice.

Messages and conversations

The metricCost per conversation that reaches a serious stage

Who it suits: Local services, clinics, and products priced by negotiation
What changes in the work: We use message objectives wired to WhatsApp, and optimise on conversations that reach a defined stage rather than on every message, so budget does not go on curiosity.

Bookings and enrolments

The metricCost per confirmed booking

Who it suits: Academies, courses, clinics, events
What changes in the work: We wire the booking system into measurement so we can see which bookings actually confirmed. Seasonality matters more than budget here: the wrong timing wastes a right-sized budget.

The technical work

What happens in the kitchen

Anyone can launch a campaign. The difference is the work behind it: tracking that is set up properly, tests that produce a result you can build on rather than someone's opinion, and decisions made against a number agreed up front rather than a feeling.

Tracking & Measurement

Why it matters: The platform learns from the data reaching it. If that data is missing or duplicated, the algorithm learns the wrong lesson and delivers your ads to the wrong people, and you pay for that lesson.
How you know it is broken: Platform numbers disagree with your actual sales, conversions record twice, or your reported conversions dropped after an iOS update.

What we do

Google Tag Manager set up with a structured container and a clean data layer instead of scattered snippets
GA4 with a consistent event taxonomy, so every event means exactly one thing
Meta Pixel alongside the Conversions API to recover what iOS and ad blockers strip out
Event deduplication with matched event_id between Pixel and CAPI so nothing counts twice
Enhanced Conversions in Google for privacy-safe customer matching
A single UTM naming convention so sources stay comparable in GA4 rather than fragmenting
Offline conversion upload from the CRM for lead businesses, so the platform optimises on the sale rather than the form
Attribution windows aligned across platforms before any comparison is drawn

Testing Strategy

Why it matters: An uncontrolled test does not give you an answer, it gives you noise. Proper structure means every result tells you a reason, not just a number.
How you know it is broken: You change several things at once and then cannot tell what moved the needle, or you kill an ad after two days and open another.

What we do

One variable per test cell: the hook, the audience, or the offer, never all three together
A clear hierarchy: campaign equals objective, ad set equals audience, ad equals creative
CBO or ABO chosen by purpose: ABO while testing so spend stays controlled, CBO for scaling
A minimum data threshold agreed before any decision, rather than a call after forty-eight hours
Creative tested in sequence: hook first because it is what stops the scroll, then body, then CTA
Audiences tested across four layers: broad, interest, lookalike, and retargeting
Holdout groups on larger accounts: we withhold ads from part of the audience so we can see what advertising actually added, not what the platform claims for itself
Landing page testing in parallel, because half the problem usually sits after the click

Optimization

Why it matters: When cost rises there is an order to the levers. Starting in the wrong place wastes a week and throws the campaign back into learning.
How you know it is broken: You reach for bids and placements first, cost keeps climbing, and the campaign re-enters the learning phase.

What we do

A fixed order of levers: creative first, then offer, then audience, and bids and placements last
Frequency and fatigue signals watched before they surface in the cost
Creative refreshed on a schedule rather than as a reaction once numbers slip
Budget reallocated by defined rules instead of daily improvised intervention
Overlapping audiences excluded so your own ad sets stop bidding against each other
Retargeting segmented by recency, because someone who visited yesterday is not someone who visited last month
Order value work (bundles, upsells) because raising value is cheaper than lowering cost

Analysis & Decisions

Why it matters: When a number is agreed in advance, the decision comes out the same whoever makes it, and you can go back later and ask why it was made. Without that number, the call changes every week depending on what looks bad on screen that day.
How you know it is broken: The report tells you what happened but not what you are doing next month, and decisions change week to week without anyone knowing why.

What we do

Daily: checking spend is tracking as expected and nothing odd has happened, without meddling with campaigns while they are still learning
Weekly: performance read split by creative and by audience, plus the cost-per-result trend
Written, agreed rules: the number at which we kill an ad, the number at which we raise spend, and the largest safe increase in one step
Customer acquisition cost across the whole account compared against LTV, not against a single order
Cross-platform overlap checked before judging any figure, because two platforms will both claim the same sale
A weekly report that opens with the return and states the next decision, rather than a wall of metrics

Which fields have you worked in, and what changes between them

The platform is the same, but what works in an online store fails in a clinic. What changes is not the ad, it is the objective itself and the number you count as success.

Ecommerce
Usual objective: Orders
Catalogue, retargeting, and order value. The biggest gain is often not acquiring a new customer, it is raising average order value and recovering abandoned carts.
Healthcare and clinics
Usual objective: Bookings and messages
Strict advertising restrictions on wording and imagery. Campaigns get rejected for reasons that are never spelled out, and the expertise is knowing how to write copy that passes and still sells.
B2B and professional services
Usual objective: Qualified leads
Long purchase cycles and small audiences. Volume is not the measure here, filtering is. Twenty right leads beat a hundred that waste the sales team's week.
Academies and education
Usual objective: Enrolments
Sharp seasonality around intake dates. Budget has to concentrate into a narrow window, and spend outside that window is largely wasted.
Technology and SaaS
Usual objective: Trials and demo requests
A precise, small audience and a high cost per click. Targeting runs on job titles and technical interests, and the content has to explain rather than dazzle.
Real estate
Usual objective: Leads
High deal values mean an acceptable cost per lead far above any other category. Filtering by area and budget saves several times what it costs.

If your category is not here, ask. If we have no experience in it we will say so in the first meeting rather than take the work and learn at the expense of your budget.

Scaling

How we grow spend without the cost running away

Scaling is not a button that raises the budget. Every spend level carries a different risk and a different job. This is the ladder we climb.

Stage 1
Validation
A limited test budget across different angles and audiences. The goal is not profit, it is learning which message and which audience respond. Leaving this stage happens on a clear signal, not on a feeling.
Stage 2
Stabilisation
We concentrate spend on what worked until cost per result settles into a narrow band. This is where the real number appears, and it becomes the baseline we measure scaling against.
Stage 3
Scaling
Gradual spend increases while watching cost move. Raising too fast throws the campaign back into learning and breaks what was built, so the steps are measured rather than jumps.
Stage 4
Saturation
Every account reaches a point where the audience is saturated and cost starts climbing. When we get there we say so and open a new lane: another platform, a new audience, a different product, rather than pushing harder in the same place.

Spend multiplying while cost stays flat: that is what gets judged. Everything else is detail.

If you are at the stage where you need to grow spend without the cost running away, this is the next step.

Request the audit

Platforms

Which platform, and when

The choice is not a preference, it follows the shape of demand in your business. On many accounts more than one platform runs together in different roles.

Meta
Facebook and Instagram. Creates demand: it reaches people who were not looking for you. Strongest for products that sell on first sight, for local services, and for message ads routing into WhatsApp. The widest reach at the lowest cost per click in this market.
Google Search
Catches demand: it reaches people already looking. Higher purchase intent and a higher cost per click. Strongest when there is real search volume around your product or around a problem you solve, which makes it excellent for services, B2B, and anything urgent.
TikTok
Cheap reach to a younger audience, and creative is everything. An ad that looks like an ad dies here. Strongest for products explained well in short video and for brands with content that is easy to shoot.
Both together
The combination that works most often: Meta creates the demand and Google catches it. When someone sees your ad on Instagram and then searches your name, both campaigns are working together even if the report attributes the sale to one.

How we work

How a week runs

Media buying is not a campaign you set up once. It is a weekly cycle of testing, killing, and scaling, and this is what it actually looks like.

Before any spend
Numbers and measurement
We work out your margin and an acceptable customer acquisition cost with you, configure the pixel and events, and confirm measurement works. Broken measurement teaches the platform the wrong thing and delivers your ads to the wrong people.
OwnerWe configure, you confirm the figures
Week 1
Testing
A set of ads across different angles, audiences and creative on a limited test budget. We test several hypotheses at once so a signal arrives quickly.
OwnerUs
Week 2
Killing
We shut down what is not working, fast. The single biggest way budgets burn is keeping a losing ad alive in the hope it improves. The call runs on a minimum data threshold agreed up front, not on instinct.
OwnerUs
From week 3
Scaling
Spend rises in steps on what proved itself, with fresh creative continuously so the audience does not saturate. Ad fatigue is real and it shows up in cost before it shows up anywhere else.
OwnerWe scale, and tell you before any large increase
Every week
The report
A short weekly report: what we spent, what came back, and where the cost went. Plus a monthly call to review the trend and decide the month ahead.
OwnerWe write it, we review it together

The account is open to you the whole time. The report saves you the reading, it does not withhold anything.

Questions

What people ask before we start

These are complete answers, not headlines. If your question is not here, ask us and you will get the same directness.

How is the fee calculated?

A percentage of monthly ad spend, agreed up front and written into the contract, and the percentage decreases as spend grows because managing 500,000 is not ten times the work of managing 50,000. The spend itself you pay to the platform directly from an account in your name, so your budget never passes through us. There is also a floor for smaller accounts, because any account takes roughly the same setup and monitoring time regardless of budget. We give you the rate and the floor on the first call, not after you have sent your details.

What is the minimum budget worth starting with?

The floor is set by the platform's need for a certain number of weekly conversions to exit the learning phase, not by our preference. The calculation runs like this: take the expected cost per conversion in your category, multiply it by the number the platform needs each week, and that is your sensible minimum. Below it the campaign keeps learning and never stabilises, so you pay the cost of experimenting without reaching the result. We work it out with your own numbers during the audit, and if it comes out short we will tell you to concentrate on one product and one audience rather than spread thin.

How long until I see a result?

The first real signal usually lands in week one or two: we learn which angle and which audience respond. That is not profit, it is information bought with a limited and clearly stated test budget. Stabilisation typically arrives within the first month, when cost per result settles into a narrow band and we know the number we are building on. Serious scaling starts from month two. Accounts with clean measurement and a good landing page move faster; accounts where we have to fix measurement first take an extra fortnight before any serious spend.

Do you guarantee a ROAS or a cost per lead?

Before seeing your account, your product and your margin, any figure quoted is a guess. After the review we give you an expected range built on cost per conversion in your category and on your own numbers, and we will tell you plainly if the arithmetic does not close. What we commit to: a campaign structure built for scale from day one, killing losers quickly rather than waiting, fresh creative continuously so the audience does not saturate, and a weekly report that opens with the return. And when the account reaches its ceiling we say so and open a new lane rather than pushing harder in the same place.

Is creative your job or ours?

We write the angles and the ad copy and specify the creative each test needs, because that is an inseparable part of the campaign rather than separate work. Production itself, the design and the video, is quoted separately or handled by your own team if you have one. In many cases the best creative is footage shot on your own phone, because audiences now scroll past anything that looks like an advert. What matters most is that creative keeps refreshing, because ad fatigue shows up in cost before it shows up anywhere else.

If I already have an account running, do you start over?

No, and doing so would waste a real asset. An account with spend history and conversion data has taught the platform something. The first thing we do is read it: which campaigns are working, which are eating budget without return, whether measurement is configured properly, and where spend is leaking. A lot of the early gains come from switching things off rather than adding things. We build on what is there and restructure what needs it, and we tell you exactly what is changing and why before touching anything.

How is this different from Paid boost in the social service?

Paid boost in the social media service is a limited budget behind content that has already proven itself organically, and its purpose is widening the reach of good content. It is not a performance campaign and it is not built on targeting, testing and conversion structure. Performance marketing is a different discipline: campaign structures built on conversion objectives, testing across audiences, angles and creative, measurement tied to revenue, and a weekly cycle that kills and scales. If you need to reach new people and sell to them at a known cost, that is performance. If you have content that works and want more people to see it, that is a boost.

What do you need from me to start?

Five things. Access to your ad account and pages, and if there is no account we open one in your name. Your real numbers: selling price, cost, margin, and average order value. That last one matters most, because without it any conversation about return is guesswork. Access to the site or a line to whoever develops it so measurement can be set up properly. Someone responsible for answering customers with a known response time, because in this market most advertising ends in a message or a call. And clarity on the monthly budget you can commit for at least three months, since anything shorter is not enough for a full test and scale cycle.

The work

Real work from this service

Each one has a page saying how it was done, the number it answered to, and where that number came from.

1,532,861 EGP in purchases across 19 days
A Meta account for a business three months old. The window is 19 days, not a full month, and the figures are exactly as Meta calculates them.
Open the project

Each one has a page saying how it was done, the number it answered to, and where that number came from.

See all the workMore work onInstagram Facebook

Next step

Start by finding out where the money is leaking

We will not sell you ad management before knowing where your problem actually sits. Send us the link and we will come back with price and timing before any commitment.

A full funnel diagnosis from tracking through to AOV: which stage is leaking, precisely
The first three things to fix, ordered by impact rather than by effort
The expected cost per result range in your category
The sensible floor for your budget so the campaign can stabilise

What happens after you send it

Within one working day
We come back with the audit price, the timing, and what we will need access to. Nothing is committed until you say yes.
Within five working days
We review the tracking, the ad account and the site CRO against your own numbers, then send the report: which stage is leaking, the first three things to fix ordered by impact, and the expected cost per result range in your category.
After that
A call to go through the report together. Act on it yourself, hand it to anyone, or work with us and the audit fee is credited against your first month.