Agency evaluation criteria: how do you see the real agency behind the pitch?

Knowing which agency evaluation criteria matter is only half the work. The other half is seeing past the pitch: an agency can prepare for anything it knows you will ask, so the evaluation has to be built around what agencies do, not what they show. This article covers the tests - which question to ask, at which stage, and how to read the answer. For the criteria themselves and how to weight them, see the full guide to media agency selection criteria.

The principle behind every test in this article is the same: score what you observed, not what you were shown. Anything an agency can rehearse measures pitch preparation. Anything you see them do under real conditions measures the partnership you are about to enter.

Testing strategic fit: create moments that cannot be rehearsed

A capabilities tour tells you what an agency wants you to know. To find out whether they understand your business, you need moments no deck can prepare for. Several techniques work; which ones fit depends on your process and the time available.

The flawed assumption. Place a deliberate error in the briefing - an audience definition that is off, a channel that demonstrably does not work for your category, a market figure that is wrong. Then watch who corrects you, and how. An agency that faithfully executes a flawed brief today will faithfully execute flawed briefs for the next three years. An agency that pushes back in writing, before the pitch, is showing you the working relationship you actually want.

The trade-off question. During the meeting, unannounced: “Your plan assumes the full budget. It has just been cut by 30 percent. What goes first, and what do you defend to the last?” An average agency shaves a little off everything. A strong agency chooses, and can explain why. The order in which things are sacrificed tells you whether they understand your business.

The live scenario. In a later round, replace part of the presentation with a working session: a real situation from your recent past, 45 minutes, solve it here. No time to polish means you see the actual team think rather than the pitch team present. Be aware that this is time-intensive with three or four agencies in the process - it belongs in the final round, not the first.

The unexpected question. The general form of all of the above: ask things no slide exists for. What you are reading in the answer is not the content but the behaviour - who speaks, who defers, who is comfortable saying “we would need to check that”.

For advertisers operating across regions, extend these tests to local market knowledge. Media consumption in Southeast Asia, for instance, differs substantially from Western markets - an agency presenting a generic global playbook is answering a question you did not ask.

Testing commercial transparency: own the format

The common advice is to request a disclosure of the agency’s income streams. In practice, a stronger approach is available: do not ask for a statement - define the structure in which every agency reports, and have them fill it in.

The reasoning is simple. When agencies report freely, you receive three or four differently framed stories, each choosing what to show. When every agency completes the same structure, every empty field is a question, and avoidance becomes visible as a gap instead of disappearing into prose.

Two conditions make this work.

Announce it early. State at the start of the process that a transparent relationship is the intended outcome, and that the financial chapters will be completed in the format you provide. Agencies then know from the beginning what they are signing up for - and how to write their proposals. A process that is explicit about transparency from round one rarely needs to disqualify anyone later; expectations do the filtering.

Close the side doors. Specify that all financial questions are answered exclusively within the requested format. Material that surfaces only in appendices is easily read past - and what is read past in the pitch tends to resurface at the contract table. If it is not in the format, it does not count.

There is a further advantage that is easy to miss: with three or four agencies completing the same structure, the pitch becomes its own benchmark. All of them want to win, so where one submission deviates from the field - a column marked “not applicable” that the others completed, a category quietly redefined - you see it immediately. Classic places to watch are agency volume bonuses and technology fees. The broader question of media transparency has been an industry issue for over a decade; the process above is how you address it before signing rather than after.

Testing anticipated rates: the too-good-to-be-true routine

Sooner or later you will face a submission that is dramatically cheaper than the field. In one pitch we organised, the leading agency came in 20% below the rest of the field. The question at that moment is not whether to be pleased but what to do next - because a gap of that size has more than one possible explanation, and they require different responses.

One possibility is that the agency wants the business badly enough to be creative with its conditions - a market-leading logo can be worth a great deal to an agency. Another is that arrangements elsewhere in the agency’s portfolio free up advantages that can be shifted towards you. Neither is misconduct, and neither is unheard of. But both raise the same question: how sustainable is it? Conditions that depend on the agency’s appetite, or on structures elsewhere, can change - and then you face fluctuation in the very numbers you selected on. A third possibility is simpler: the promise was made to win and cannot be kept.

You cannot always tell which scenario you are looking at from the submission alone. The routine that separates them:

Take two agencies into contract negotiation and go deep. Have the low submission explained: which volumes, which inventory mix, which deal structures does it rest on? Does the story hold?

Make the rates contractual. The submitted rates become an annex to the contract, with compliance checks attached. How far an agency is willing to go in committing is itself the answer: an agency that meant it, signs. An agency that did not, becomes vague.

Decide with open eyes. If it still does not add up, choosing the number two is often the better outcome. If you proceed anyway, do so knowing what you are accepting - this is a judgement to make together with your leadership, not a line item to score.

Testing accountability: agree the key points during the pitch

Accountability is genuinely hard to assess from a presentation - every agency says it measures, reports, and owns its results. Objective anchors exist (whether an agency is SOX-compliant, for instance, tells you something about its internal controls), but the more revealing test is one you can build into the process itself.

Reach agreement on the key contract points before a winner is chosen. Depending on the structure of your pitch, this happens in the first or second round: the principles that matter most - transparency provisions, audit rights, data ownership, the measurement framework - are agreed in outline while several agencies are still competing.

This does three things at once. It prevents surprises after the decision, when your leverage is reduced. It manages expectations on both sides early. And it lets you observe how each agency responds to accountability requirements while it still has something to lose - which is the most honest measurement available.

One caveat: this is not a form-filling exercise. Reading those responses correctly requires experience at the table - a human in the loop, not a checklist.

For translating what you observe into comparable scores across the panel, the agency scorecard guide covers the mechanics of scoring and calibration.

Testing the team: make the account team a pitch condition

The concern is familiar: senior people present, other people deliver. The test is to remove the ambiguity before it can arise.

Write it into the pitch conditions: the account team presents. Not as a preference - as a rule of participation. If it becomes clear during the presentation that a pitch team is at the table instead of the account team, participation ends. Whether that is said in the room depends on the setting and how clear the situation is, but the rule itself should be unambiguous.

This is not about excluding leadership - the opposite. You want the directors at the table: they open the meeting, present the credentials, introduce their team, close the conversation. They are the people you will negotiate the contract with, who join at strategic moments during the year, and who can take decisions above the team. That relationship is worth building from the first meeting.

The distinction is about roles. The moment the conversation turns to the specifics of your account, the account team should be doing the talking - that is when you see them work. A director who answers the account questions while the team looks on is not breaking a rule, but it is a signal, and it belongs in your scoring.

One extension worth writing into the conditions: replacing named team members after appointment requires the advertiser’s consent. The team that won is the team that stays - or you co-sign the change.

The pitch itself is the biggest test

Everything above shares one logic: the behaviour of an agency during a selection is a preview of the behaviour you will experience as a client. An agency that responds to clarifications promptly, asks sharp questions, corrects your assumptions, completes your formats without manoeuvring, and commits its numbers to the contract is showing you the next three years. How an agency negotiates before appointment is almost always how it behaves when the agreement is tested.

None of this requires treating agencies as suspects. Most agencies that reach a serious shortlist are professional organisations, and the strong ones tend to welcome a well-structured process - it lets quality show. The tests exist because a pitch is the one moment when everything is observable and everything is negotiable. Using that moment well is simply good governance.

If you are still deciding whether a full pitch is the right intervention for your situation, the comparison of pitch, renegotiation, and audit as distinct options may help clarify which tool fits your circumstances.

Frequently asked questions

What questions should you ask an agency during a pitch? The most revealing questions are the ones no slide exists for: where the agency would not invest your budget, what it would cut first if the budget dropped, and what assumptions its rates rest on. The value is rarely in the answer itself but in the behaviour - who speaks, how the team handles pressure, and whether they are comfortable admitting what they do not know.

When should commercial basics be discussed in a selection process? Earlier than most advertisers expect. Announce the transparency expectations at the start and reach outline agreement on the key contract points while several agencies are still competing. Once a winner is chosen, your leverage starts to drop.

How do you know whether a very low rate submission is realistic? You test it rather than guess. Take the agency into contract negotiation, have the underlying volumes and assumptions explained, and make the rates a contractual annex with compliance checks. An agency that meant its numbers will commit to them; one that did not will become vague. If it still does not add up, consider your options.

Who should present in an agency pitch? The account team - the people who will run the business day to day - and this should be a written condition of the pitch. Leadership is welcome and valuable: opening, credentials, introductions, and the contract conversations. But when the discussion reaches the specifics of the account, the account team should be doing the work, because that is who you are actually selecting.