A Marketing Agency RFP That Tests Real Work, Not Pitches
Most agency RFPs and quarterly reviews compare the wrong things: polished capability decks and case studies pulled from a completely different industry, instead of a real test of whether the agency can do the specific work you need at your scale.
The fix isn't a longer questionnaire. It's asking for a small piece of real work, tied to your own data, and holding an existing agency to specific numbers you agreed on before the engagement started. Both problems come from the same root cause: judging a vendor on how they present rather than on what they'd actually produce for your account.
What to ask for instead of a capabilities deck
Give two or three finalist agencies the same brief and a short, fixed window, and ask for an actual work sample against your own industry and data rather than a generic case study. Ask specifically who will work on your account day to day, not just the partner who runs the pitch, and how many other accounts that same team currently handles, since a team stretched thin produces worse output regardless of how strong the agency's brand is. If an agency won't share the working team's actual account load, treat that refusal as information in itself.
Red flags that show up in the first call
Watch for a handful of patterns:
- Insisting on a long contract before doing any diagnostic work on your account. - Case study results with no comparable starting point you can sanity-check. - A named pitch team that's different from the proposed working team. - No clear, specific answer for how they'll report performance to you week to week. - Reluctance to name any account they've lost or a campaign that underperformed, since an agency with no honest failure story to tell is usually one that hasn't been challenged yet.
Reviewing an agency you already have
Before blaming an existing agency for weak results, rule out your own gaps first: broken landing page tracking, stale list hygiene, or a lead-routing problem can all suppress numbers that have nothing to do with the agency's work. Ask for raw channel data, not just the summary they present, and if they're running paid search for you, ask for actual cost per click by campaign rather than a single blended number, and compare it to the broader market average of around $5.261 as a sanity check, not a hard target. A campaign running far above that on cost per click isn't automatically bad, but it deserves a specific explanation, not a shrug.
Setting a real bar before the engagement starts
Agree in writing on what specific numbers count as success at 90 days and at 6 months, tied to a pipeline stage you both consider real, like a sales-qualified opportunity, not just a form fill. Decide in advance what happens if the agency misses that bar for two review periods in a row: a scope renegotiation, or an exit, rather than an open-ended conversation once the numbers actually come in short. Put the same bar in writing for a brand-new engagement, so nobody is negotiating the definition of success after the first disappointing month, and revisit the bar together at the 90-day mark rather than letting it quietly become whatever the current run rate happens to be.
What a fair diagnostic brief actually looks like
Keep the paid diagnostic narrow enough that a serious agency can turn it around in one to two weeks: a real segment of your list, a real product or offer, and a specific question, like how they'd structure the first 30 days of outreach to that segment. Score the response on the same rubric across every finalist, covering the actual strategy, how well it reflects your business specifics rather than a generic template, and how clearly they explain the reasoning behind it, not on production polish alone. Share the same rubric with the finalists up front, since an agency that pushes back on being scored this way is telling you something about how they'll handle an honest performance conversation later.
What Good Looks Like
A disciplined agency evaluation tests real work against your own data and industry rather than polished case studies from another vertical, and locks in specific 90-day and 6-month performance numbers, tied to a pipeline stage you both consider real, before the engagement starts.
Building The Capability (5-Stage Skill Ladder)
How to Get Started
Frequently Asked Questions
How many agencies should we include in a formal RFP?
Two or three finalists is usually enough for a real comparison. Beyond that, you mostly spend time reviewing pitches instead of testing actual deliverables, and the marginal finalist rarely changes the outcome of the decision.
Should we pay agencies for RFP work?
Yes, if you want a genuine work sample instead of a recycled pitch. A small paid diagnostic filters out agencies unwilling to do real work before a contract is signed, and it gives you something concrete to compare across finalists.
What's a reasonable review cadence for an existing agency?
Short monthly check-ins on leading indicators, with a deeper quarterly review against the 90-day and 6-month numbers you agreed to at the start. Waiting for an annual review to raise a performance concern usually means you've already lost months you can't recover.
Is a lower monthly retainer always the better deal?
Not necessarily. Compare the retainer against the media spend the agency will manage and the actual seniority and hours included. A cheaper retainer staffed by junior team members with high turnover often costs more once you account for rework and lost time.
Sources
Where we quote a benchmark, we show its source. Other figures in this guide are estimates or general guidance, so check them against your own numbers.
- Google Ads Average CPC (All Industries). WordStream 2025 Google Ads Benchmarks, 2024-2025.
Related Guides
Where B2B Conversion Rates Actually Fall Short of Plan
Why comparing your funnel to a published conversion benchmark usually misleads you, and what to check on your own numbers instead.
Why Your Webinar Registrant List Isn't the Same as Your Pipeline
A checklist for tracking field events and webinars past the registration count, so you can tell which ones actually produced pipeline and which didn't.
When Content Syndication Pays Off, and When It's Just Expensive Leads
How to judge whether a content syndication network will bring in usable pipeline or just a list of names that never respond to a follow-up call.
What Actually Belongs on a CMO's Executive Dashboard
A short audit for cutting a bloated marketing dashboard down to the handful of numbers leadership actually needs to see.
Owning the Search Results for Your Category's Highest-Intent Terms
A practical way to identify and win the small set of search terms that show up right before someone buys, instead of chasing broad traffic volume.
How to Audit Your B2B Demand Gen and ABM Program This Quarter
A step-by-step B2B demand gen and ABM audit you can run in an afternoon with data you already have: pipeline, scoring, campaign ownership and message match.