Retainer vs Project: Which Model Actually Fits.

This question is usually argued as a pricing question, which is why it is usually answered wrongly. Retainer and project are not two prices for the same thing. They are two shapes of work, and the choice is decided by a single test: can you write down what finished looks like?

If you can — an identity, a website, a launch, a film — it is a project. If you cannot, because the work is a process rather than an object, it is a retainer. Everything else follows from that.

The core difference

When a retainer is the only honest answer

Some work cannot be packaged with an end date without damaging it. Paid advertising is optimisation — the optimisation is the work, and it stops when the retainer stops. Social media depends on rhythm, and a gap costs more than the gap itself because the audience and the algorithm both notice. SEO compounds, and compounding is the whole return.

The clearest illustration we can give from our own accounts: in the ICAR campaign the cost per conversion held at 4.87 EGP across 2,235 conversions while budget scaled. That did not happen in a month, and it did not happen by itself — it happened because new creative shipped every week. Stop the weekly work and the cost climbs within a fortnight.

When a project is the right call

• Brand identity. A clear deliverable, a handover, and nothing to optimise afterwards.
• A website or app. The same, with maintenance as a separate, smaller arrangement.
• A launch campaign anchored to a date.
• A shoot. Defined days, defined outputs.
• An event. One night, one scope.
• An audit. A finite diagnosis, whether or not the treatment follows.

The SEO case, because it splits

SEO is where this question is asked most often, and it is genuinely both. The first phase is a project: a technical audit, the fixes, the structure, and the first set of pages. It has a scope, a cost and a finish line, and a good agency should be willing to sell exactly that.

What follows is not a project. Content production, internal structure, authority building and monitoring compound over months, and the moment they stop, competitors who did not stop pass you. Any agency that sells SEO as a one-off deliverable is selling the first phase and letting you believe it is the whole thing.

Our own model reflects that split: SEO management at 20,000–37,000 EGP a month is the ongoing part, and the audit that precedes it is priced separately. A fuller discussion of what the retainer should contain is in choosing an SEO agency.

Performance-based pricing, honestly

The third model that comes up is paying per result, or a percentage of what the work produces. It is attractive because it looks like shared risk, and it works in a narrow set of cases: when results are cleanly attributable, when the agency controls enough of the funnel, and when the volume is high enough to be statistically real.

It fails in most service businesses, for a reason that has nothing to do with the agency: the client's own sales process decides whether a lead becomes revenue. Paying per lead then rewards volume over quality, which is the opposite of what a real estate developer or a clinic needs. We have published figures that make the point — a qualified real-estate lead at 10.66 EGP outperformed cheaper leads precisely because the filter happened before the lead, not after.

The sequence that protects you

With an agency you have not worked with, start with a project. A website, an audit, a shoot, a single campaign — something with a scope and a date. In six weeks you learn the three things a pitch cannot tell you: whether the work is good, whether they answer, and whether they hit a deadline.

Then move to a retainer if the answer is yes. That sequence costs slightly more in total than signing a long agreement up front, and it is cheaper than every bad year-long contract ever signed.

Three mistakes that cost real money

1. Running ongoing work as repeated projects. Commissioning a 'campaign' every two months costs more in total and throws away the account's accumulated learning each time. The second campaign starts from zero even though it should start from the first one's data.
2. Signing a twelve-month retainer untested. A long commitment with no exit clause does not buy loyalty; it removes the agency's reason to keep earning the month.
3. Buying a retainer for work that finishes. Paying monthly for a website that was delivered in March is the quiet version of this, and it usually survives because nobody reviews the line item.

What a fair retainer contains

Whichever model you choose, the contract should answer five questions in writing, because every dispute we have ever been asked to clean up traces to one of them going unanswered:

• Scope. How much production is included, how many revisions, what counts as extra.
• Response time. Not 'promptly' — a number.
• Reporting. What is reported, how often, and in whose account the data lives.
• Ownership. Ad accounts, content, source files and domains in your name from day one.
• Exit. Notice period, what you receive on exit, and when.

The clause-by-clause version is in what belongs in an agency contract, and the ownership question specifically in who owns your accounts when you leave.

A one-line test

Before signing anything, write down what you will be able to point at in ninety days. If the answer is an object — a site, an identity, a film — you want a project. If the answer is a number that should be better than it is today, you want a retainer, and you should agree now how that number will be measured.

Our prices for both shapes are published on the pricing page rather than quoted on request, and if you are deciding which your situation needs, describe the work — if it is a project, we will say so instead of selling a retainer.