Fractional CMO · 90-Day Plan
Prepared for Mahendra · The Monk · SPF Foods · August 2026
Proposal · Fractional CMO · 90 days
Measurement you can read weekly. Guests you can identify. A channel you can reach them through without paying a commission. And a written procedure, so branch six inherits all three on the day it opens.
Ten steps, in dependency order. Each one states what gets built, what it produces, what it depends on, and how you know it worked.
01 · Where things stand
From our call and from what is publicly visible. If anything below is wrong, it changes the plan — so it is worth correcting before we go further.
On the ticket falling while volume rises. That pattern normally means the same guests are ordering more often, in smaller and more solo occasions. It is a base, not a decline — a weekday habit is the hardest thing to build and you already have it. The work is to raise what a solo order is worth (step 05) and to add the larger occasions on top of it (steps 06 and 09), rather than to discount your way back to bigger tables. Whether that reading is right is the first thing step 01 tests.
02 · What the plan is actually building
Marketing activity is not the point. These four are the point. Every one of the ten steps builds exactly one of them, which is how you can check the plan rather than take my word for it.
One sheet per branch, every Monday, that you and five managers read the same way. Without it, nothing after it can be proven — which is the position the last four attempts were judged from.
Built by steps 01 · 02Knowing who ate, at which branch, through which channel, and when they last came. Today a guest of fifteen years is indistinguishable from a first-timer.
Built by step 03A way to contact those guests that you own, that costs almost nothing per message, and that does not disappear when an ad budget stops.
Built by steps 04 · 06 · 07The above, written as a procedure with a checklist and an owner, so opening a branch is a process rather than a fresh improvisation each time.
Built by step 10Steps 05, 08 and 09 are revenue work — ticket, local search and catering. They sit on top of A to D, and they are deliberately scheduled after them, for the reason set out in section 04.
03 · The plan
Each step is a thing that exists at the end of it. “Done when” is written so that you, not I, decide whether it happened.
04 · Why this order
This is the part that makes it a system rather than a list of marketing activities. The sequence is not a preference — reordering it breaks it.
First — nothing above this is provable without it
Then — these need a measured baseline and a named guest
Then — new demand, once there is somewhere to keep it
Last — only writable once the rest has actually been run
05 · The arithmetic
I have not put a total return on this page, because any total I produce before step 01 is a guess with a decimal point on it. What follows is arithmetic you can check, and one illustrative scenario with every input visible.
| Per 58 AED order | You collect | Cost of sale | Contribution |
|---|---|---|---|
| Dine-in | 58.00 | 17.40 | 40.60 |
| Direct delivery | 58.00 | 24.77 | 33.23 |
| Aggregator | 40.60 | 18.90 | 21.70 |
| The gap, direct vs platform | — | — | 11.53 per order |
Modelled at 30% food cost, a 30% blended platform take rate, 1.50 packaging, 5.00 own-fleet fulfilment and 1.5% payment fees. Published UAE commissions run 15–35% before service fees and placement costs, and they differ per platform and per branch. These are my assumptions, not your figures. Step 01 replaces all of them in week one, and the numbers below move with them — in either direction.
| Lever | One unit is worth | Units per month to cover 12,000 | Per branch, per day |
|---|---|---|---|
| A delivery order moves off a platform | 11.53 | 1,041 orders | 6.9 |
| One side or drink added to an order (+8 AED) | 5.60 | 2,143 orders | 14.3 |
| A quiet guest orders once more | ~33.00 | 364 orders | 2.4 |
| An office account ordering weekly at ~1,200 | ~2,640 / mo | 4.5 accounts | — |
Read each row on its own and decide whether it looks achievable across five branches. That judgement is worth more than any projection I could write, because you know the floor and I do not. The plan runs all four levers, so in practice they share the load rather than any one carrying it.
| Gross profit contribution | Month 1 | Month 2 | Month 3 | Assumption behind it |
|---|---|---|---|---|
| Orders moved off platforms | — | 4,600 | 10,400 | 400 then 900 orders a month, across five branches |
| Ticket, via sets and attach | — | 7,400 | 13,900 | 8% then 15% of orders take a +8 AED addition |
| Quiet guests returning | — | 9,900 | 26,400 | 300 then 800 orders from the messaging in step 06 |
| Catering accounts | — | 7,900 | 18,500 | 3 then 7 accounts, ordering weekly |
| Month total | 0 | 29,800 | 69,200 | Month one installs; it does not harvest |
| Ninety days | 99,000 gross profit against 54,000 of spend — 36,000 fee plus roughly 18,000 media | Net positive by about 45,000 | ||
This scenario is deliberately modest and it is built to be missed visibly if it is wrong. Every input in the right-hand column is a line on the weekly sheet from week two, so by week six you will know which way it is going rather than finding out at day ninety. It assumes roughly 16,500 orders a month across five branches — my estimate, and one of the first things step 01 corrects. Exiting at a month-three rate near 69,000 a month against a 12,000 fee is the case for continuing; failing to reach it is the case for stopping, and section 09 says how.
06 · Limits
You have paid for confident forecasts before. The useful thing I can offer instead is an honest account of where this plan is weak, so you can judge it on something other than tone.
07 · How the work runs
One reading of the sheet, and decisions where a decision is needed. Plus roughly twenty minutes a week from one nominated person at each branch, which is the real dependency in step 03.
Not an agency of record, and not a media buyer taking a percentage of your spend. Media is bought on your accounts, at a number you approve. Print, software licences and any production costs sit outside the fee and get quoted before they are incurred.
08 · What I need from you
09 · Terms
12,000 AED per month. Invoiced monthly in advance. No setup charge, no success fee, no percentage of spend.
Media is proposed in week two, from the audit, with the reasoning attached. You approve the number and it is spent on your accounts. Print and software sit outside the fee and are quoted first.
Three months. Month one commits, because a half-finished installation is worth less than none. Months two and three run month to month.
We read the weekly sheet together. If it has not moved, we change the plan or you stop. Written into the agreement, not offered verbally.
Every account, list, asset and document is created in The Monk’s name on day one, including the contact list and the playbook.
Either you run it with the playbook, or we agree a lighter arrangement to keep operating it. There is no renewal built into this proposal.
Next step
Before agreeing to anything, the useful next hour is opening the POS and the aggregator dashboards and replacing the estimated figures in section 05 with real ones. If they hold up, the plan stands. If they do not, you will have found that out for the cost of an hour rather than a quarter.