Big outcomes.
Bigger thinking behind them.
Impossible briefs. Precise decisions. Results with a story worth opening.
From the practice of Yousef AbdoFounder & Architect ↗Get inside the brief. Follow the decisions. See what changed.
11 / 11 files
SkincareCase study
GiftingCase study
FragranceCase study
E-commerceCase study
FragranceCase study
E-commerceAccount note
Lead generationAccount note
E-commerceAccount note
E-commerceAccount note
E-commerceAccount note
E-commerceAccount note
Every number belongs to a specific account and period. Open a file for the decisions and context behind it.
The performance register. Individual account results, with their own currency, period and context.
Select a channel to explore ↓META-01
662 purchases at SAR 21.30 each. SAR 14,100 of spend returned SAR 100,913. Read the campaign column, not the average: some rows sit at 8.72, some at 5.77, some at zero. The 7.16 average is not because everything works — it is because what does not work gets cut early.
META-02
The whole funnel: content views, adds to cart, payment info added, purchases. 2,942 views became 361 carts, 158 payment details and 100 purchases worth SAR 31,394. These columns are not decoration — each step tells you where the leak is, and the gap between 158 and 100 is a checkout problem, not an ad problem.
META-03
1,486 purchases; SAR 31,255 of spend against SAR 229,853 in sales. Campaigns here run between 4.97 and 12.44, which is normal — a healthy account is not uniform, it is one where the weak line is constantly removed and the budget moves to the strong one.
META-04
The strongest Meta account in this set: SAR 6,096 of spend returned SAR 70,378 across 449 purchases at SAR 13.58 each, with one campaign alone at 12.80. A number like this does not come from clever targeting — it comes from the offer being right before the ad ever ran.
META-05
An Egyptian account in pounds: EGP 141,038 of spend against EGP 1,197,901 in sales across 3,445 purchases, with an average basket of EGP 347.72. The AOV column is the most important thing on this screen — it sets the maximum you can afford to pay for a customer, and without it the return is a number with no ceiling.
META-06
A messaging campaign rather than a direct-purchase one: 770 conversations at EGP 12.58 each, from 80,102 reach and 393,111 impressions. It is here because not every business sells on a website. The conversation is the first step of the sale, and what gets judged after this is the close rate, not the conversation count.
GOOGLE-01July 2024
One month: SAR 15.1K in cost against SAR 288K in conversion value at a 3.69% click-through rate. Google returns numbers like this when the demand already exists before you pay — people are searching the product by name. That is not clever targeting, it is the right channel choice.
GOOGLE-02Sept 2023 — Jan 2025
Sixteen months on one screen: SAR 22.4K in cost against SAR 167K in sales. The chart shows the seasonality plainly — some months go to zero. An account judged on a single month is judged wrong, and this is exactly the picture that shows why.
GOOGLE-03April 2024 — April 2025
1.57M impressions at an average cost per click of SAR 0.32, SAR 18.9K in cost against SAR 135.1K in conversion value. Notice the line stays flat for months before it moves — that period is negative-keyword cleanup and catalogue feed work, and that work is why a click costs 32 halalas.
GOOGLE-04
SAR 24.6K in cost against SAR 480K in conversion value at a 3.39% click-through rate. The highest return in this whole set, and it is Google rather than Meta — because this account was selling to people already searching for it. Channel choice before campaign optimisation.
GOOGLE-05March 2025
A smaller month: SAR 4.6K in cost against SAR 28.4K in conversion value, 121K impressions, clicks at SAR 0.32. The small account sits beside the large one deliberately — the same method works on a small budget, and the difference is scale, not approach.
GOOGLE-06
The comparison that matters: same account, Search against Performance Max. Search produced 223 purchases on SAR 11,496 at a 6.17% interaction rate, and a second search campaign hit 22.07%. Performance Max produced 34 purchases on SAR 3,010 at 1.75% from 894K impressions. Volume of impressions is not an achievement.
TIKTOK-01
410 completed payments at an average of SAR 241.12, totalling SAR 98,858. The important thing on this screen is the spread: campaigns at 5.05 sitting beside campaigns at 0.52. That is how TikTok works — you test cheaply, kill the losers fast, and move the winner to Meta with a bigger budget.
TIKTOK-02
The TikTok funnel shows 389,495 page views, 111,400 content views, 31,792 adds to cart, 33,362 checkouts initiated and 1,635 completed payments worth SAR 432,315. Checkout events exceed add-to-cart events in this capture. Review the event definitions and reporting window before treating the columns as a sequential conversion funnel.
TIKTOK-03
438 purchases at SAR 16.76 each, SAR 7,339 in cost against SAR 62,772. Three rows on this screen show zero sales, and one burned SAR 46 for nothing. Those rows are not failure, they are the cost of knowing — without them you never find out where the 11.31 came from.
TIKTOK-04
972 purchases at an average cost of SAR 20.92 and a value of SAR 141,651. One campaign alone produced 245 of them; another produced a single purchase at SAR 100.75. The work here is not improving the weak campaign, it is closing it and moving its budget to the one that works.
SNAPCHAT-01
$287,113 in tracked purchase value from a single Snapchat account. Snapchat gets dropped from most plans for not being a "serious" platform — which is exactly the opportunity: lighter competition and cheaper reach for the same Saudi segment.
SNAPCHAT-02
Six auto-bid ad sets, $6,789 of spend in total. The split is deliberately uneven: one set took $2,001 and returned 165 purchases while another took $180 and returned 4. Budget follows performance; it is not shared out equally.
SNAPCHAT-03
7.1M impressions, 74,866 clicks at $0.37 each, and 1,849 purchases worth $166,386. The eCPM sits at $3.84 — materially cheaper than Meta for the same audience, and that is the entire reason Snapchat is in the plan.
SNAPCHAT-0412 — 30 June 2025
Eighteen days: $1,695 of spend against $14,358 in sales from 353 purchases. The chart is flat at zero until 16 June — this is a launch from nothing, and reaching 8.47 inside a fortnight happens when the offer itself was right before the ad ran.
SNAPCHAT-0512 June — 10 July 2025
The same account after scaling: $4,966 of spend against $32,260 and 780 purchases. The return fell from 8.47 to 6.5 as spend tripled — that is not a decline, it is the normal price of scale, and not knowing that is how people kill a campaign that was still profitable.
SALES-01
The most important shot on this page. Gross sales of SAR 1,209,889 — and after coupons (SAR 466,114), product cost, shipping and payment fees, the net is SAR 659,663. Nearly half the number disappears. A media buyer shows you the first figure; I hold myself to the last one.
SALES-02December 2024
A month at 1,881 orders and SAR 293,469 net. Shipping alone consumed SAR 36,654 — more than ten times the coupons. Any Saudi account that leaves shipping out of its margin maths is reading a profit that is not there.
SALES-03November 2024
3,037 orders and SAR 444,903 net, up 85.66% on the prior month. The spike on 10 November is not luck — it is a seasonal campaign prepared a month ahead. Season planning happens while you are still outside the season.
SALES-04October 2024
1,777 orders and SAR 239,917 net. Look at 7 October: a sharp one-day spike that returns straight to baseline. That single day moves the monthly average on its own, which is why judging an account on a daily mean misleads you.
SALES-05Jan — May 2025
The bad number, and it is here on purpose. 1,005 orders with net down 54.95%. The cause is right there on screen: coupons ate SAR 110,969 out of SAR 339,844 — a third of sales went to discounting. A discount is not a growth tool, it is a loan against your margin that you repay every month.
SALES-06
The dashboard reports SAR 420,202.22 in gross sales and SAR 419,860.46 in net sales, with no coupons shown. Shipping and payment fees appear as separate lines. Read the dashboard totals as reported; this capture alone does not reconcile every line into a profit calculation.
SALES-08July — December 2024
Six months: 2,262 orders, SAR 870,045 gross, SAR 466,408 net. Coupons alone accounted for SAR 355,145 — 41% of gross. And the chart shows orders declining from October to December. The two are connected: once you train a market on discounts, it waits for the discount.
SALES-09
The store dashboard reports SAR 1,327,100 in sales, 6,581 orders, 22,260 visits and a displayed conversion rate of 3.3%. These are the figures shown in the capture; the conversion figure is not calculated from the displayed order and visit totals.
SALES-102025
A full year: SAR 1,248,874 from 4,710 orders. Three clear peaks in March, May and September, and a near-flat line in between. This account lives on its seasons, which means budget has to move with the peaks rather than being spread evenly across twelve months.
SALES-11
2,274 orders worth SAR 725,168 — an average basket around SAR 319. Average basket sets the ceiling on acquisition cost, and anyone who asks me "what ROAS?" before telling me their basket size is asking the wrong question.
SALES-12February 2025
A small store: 277 orders worth SAR 77,087 in a month. It sits beside the million-riyal accounts because the same method works at both sizes. The difference is that here every decision shows up faster, because there is no volume to hide a mistake behind.
SALES-13
The lifetime picture for one store: SAR 5,273,829 in sales, 16,707 orders, 404,839 visits, converting at 4.10%. Conversion is rising while visits fall — that is the signal that the traffic got cleaner, and it is worth far more than visit growth.
Founder / Four source markets
Market notes from Yousef Abdo’s personal portfolio.
Cash on delivery makes the placed order a meaningless number. The real number is the order delivered and collected, and the gap between them can be a third of the budget. Any account here that is not tying delivery data back to campaigns is measuring a fiction.
2023 — nowThe most expensive market in the region, and the one most forgiving of a high price and least forgiving of a trust gap. Acquisition cost runs several times Egypt, so the account that wins is the one where customer value is modelled over a year rather than an order. Instalments are not a discount; they are an adjustment to the CAC you can afford.
2024 — nowA market that moves slowly and rewards patience. Buyers compare, ask around, and come back a week later, so a campaign judged at seven days is judged wrong. The attribution window here needs to be at least twenty-eight days or you will kill campaigns that were working.
2025 — nowSmall and expensive, and the audience pool burns in a fortnight. Creative rotation here is not an optimisation, it is a survival condition. And because the pool is finite, growth comes from raising customer value rather than from reaching new people.
2025 — nowWhen the operating system needed to improve, the work became the product.
Agencies run media buying teams on spreadsheets and WhatsApp groups. Nobody knows who is on what, or which account needs a hand today. So I built the system I needed.
Manual prospecting eats hours a week and produces unstructured output. So I automated the boring half and left the judgement to a human.
Original ad creative and a session from the practice.
Client relationships / 34 marks
Brands from Yousef Abdo’s client portfolio.
Relationships built through the work.
Our members have worked across nine countries. Experience carried into the collective from different corners of the world.
In focus / EG01 / 09
The collective’s shared experience. The case files above document individual projects and their own results.
The work sets the bar.
Meet the people behind it