How to Scope a Paid Agency Trial That Can Actually Fail
You have decided to pay an agency for a trial before you sign a retainer. A loose trial brief lets any competent agency pass, and then you sign on a result that could not have gone the other way.
A trial every agency passes is a formality
The loose version of a trial brief is one sentence: run our social for a month and we will see how it feels. It measures how pleasant the agency is and how well its best people perform while the pitch is fresh, not how the account will run in month four.
A trial only tells you something if it is built to surface the specific risk you are worried about, and it must be possible to fail it. This post assumes you are already going to buy a paid pilot before a twelve month retainer. It is about designing the scope, not about whether to buy one.
Start from the risk you are worried about, then pick the test
| The worry | The trial design that surfaces it |
|---|---|
| They will be slow or miss dates | A fixed delivery date and time, with a real downstream commitment such as a launch the agency is told about. |
| The team that pitched will vanish | A named account lead and producer written into the trial scope. |
| They will not take feedback | One built-in revision round with a stated turnaround. |
| They do not really know our category | A deliverable that cannot be templated from another client, so the agency must ask you questions to finish it. |
| Reporting will be vague | A small reporting deliverable due on a fixed day, judged on whether it arrives and covers what was agreed. |
Pick one or two worries, not all five. A trial that tests everything tests nothing, and it turns into a small retainer. Every row is about process and reliability, not whether the work is good.
Ask for something that ships, not a strategy deck
A strategy deck can be written by the agency’s best writer over a weekend and says nothing about who will run your account. A shipped deliverable forces the agency to run its actual workflow: asking you for assets, routing approvals, hitting a date.
Here is a hypothetical. A regional home-services brand asks two agencies for a two-week batch of scheduled posts, published through the brand’s own approval steps. On day two the first agency asks for missing assets and flags that one manager reviews everything. The second delivers a deck about content pillars and asks no questions. Only the first has shown how it behaves when something is in its way.
The good version is bounded, small enough to finish inside the trial window, real enough that a delay would cost something, and touches the tools and people you will have after signing. The bad version is open ended, a document only, or work on a fake brief no one will use.
The same trial brief, written two ways
Vague brief, hypothetical home-services brand. Run our social for a month and we will review at the end.
Scoped brief, same hypothetical brand. Deliverable: one two-week batch of scheduled posts for the spring booking campaign, taken through our approval steps and published by the agency. Delivery: full batch submitted by 10:00 on Thursday, ahead of the campaign launch the following week, which the agency has been told about. People: [account lead name] and [producer name], both on every call. Revision: one round, returned within two working days. Our side: assets by day three, one approver. Pass criteria: agreed in writing before day one.
The named person directly tests whether the people who pitched are the people who deliver, because the team that pitches you is not the team that runs your account. If the agency will not name a person for the trial window, treat that as the result, not a scheduling detail.
The deadline must carry a real cost of missing it, such as a launch or an internal date the agency knows about. Otherwise a slipped date is only an anecdote.
Write the pass criteria before the trial starts
Criteria you can check without any opinion about taste:
- Delivered by the agreed date and time.
- The named person attended the calls and answered questions.
- Questions and asset requests came early, not at the deadline.
- The revision round came back inside the agreed window.
- Anything unclear was flagged instead of guessed.
This post does not judge whether the work is good. Quality is a separate question, so keep it out of pass or fail. These criteria test whether the agency did what it said it would do.
Scope your own side too. Approval turnaround, asset delivery and access are your job, and a slow approver makes any agency look bad. Check your own approval process first.
Agree the criteria with the agency in writing before day one, and decide now what counts as a pass, an extension and a fail.
Pay for it, paper it, and decide before you see the result
The trial needs its own short scope document: a stated fee, the deliverable, the date, the named staff and the criteria. Our guide on how to read an agency scope of work covers what to look for.
Three things get forgotten, and all three belong in writing. There is no automatic conversion into a retainer. Name who holds the accounts and access during the trial and how it is removed at the end, as in our post on who owns your social accounts. And settle who owns the work product if you do not continue. None of this states what the law provides, so have a lawyer review the trial scope if you want certainty.
Then write the decision rule. A pass converts into the same named people carried into the retainer scope. A fail means you stop. One extension is allowed, and you write down what it would need to change. Deciding after you see the output invites rationalising it. A trial designed so that the agency cannot fail is only a small first invoice.
Next step
Browse vetted agencies to build your shortlist, or, if you run an agency, list your agency free.
FAQ
How long should a paid agency trial run?
Length matters less than fit. Make the window just long enough for one real deliverable to ship with one revision round.
Should the trial use a real brief or a made-up one?
Real. A fake brief lets the agency skip asset requests, approvals and dependencies, which is where the workflow shows. Keep it small in size and real in consequence.
Sources
- Buy a Paid Pilot Before You Buy a Twelve Month Retainer, fetched 2026-09-20
- How to Read an Agency Scope of Work Before You Sign It, fetched 2026-09-20
- Who Owns Your Social Accounts When an Agency Runs Them, fetched 2026-09-20
- The Team That Pitches You Isn’t the Team That Runs Your Account, fetched 2026-09-20
- Your approval process is shaping the work more than your brief is, fetched 2026-09-20
