How to choose a demand generation agency without buying activity

Read enough agency contracts and you start noticing they are all written in the same grammar. Four campaigns a month. Two landing pages. Twelve creatives. A monthly reporting call. It reads like a menu because it is one, and the reason is simple: deliverables can be counted, invoiced and defended, and outcomes cannot be promised without risk.

That is a rational way to sell a service. It is a terrible way to grow a company, because once the contract is written in outputs, the whole operating rhythm reorganises around producing them.

What the activity contract actually buys

Consider what happens in month three of a typical engagement. Performance is flat. The honest move is to stop two of the four campaigns, put the money behind the one that is working, and spend the freed time rewriting the offer. The contracted move is to ship four campaigns, because four campaigns is what the client is paying for and cutting two looks like under-delivery.

Nobody in that room is acting in bad faith. The account manager is protecting the renewal. The client is protecting the budget line they defended internally. The scope document is the only party with an opinion, and its opinion is that motion equals value.

When you pay for output, you get output. It will arrive on time, look professional, and change nothing.

How to choose a demand generation agency: the three tells

• There is a layer between you and the person doing the work. Every question takes a day to answer because it travels through a translator who cannot make a decision.
• Reports lead with volume. Impressions, sessions, MQLs, content shipped. Pipeline appears on slide nine, if at all.
• Nobody has ever recommended you spend less. An agency that has never argued against a budget increase is not a partner, it is a vendor with a quota.

The alternative is not heroic, it is structural

You do not fix this with a better agency. You fix it by changing what the contract counts. Three structural changes do most of the work.

1. 1

   Contract a small number of outcomes, not a long list of outputs

   One or two metrics that the business already cares about: qualified pipeline created, cost per opportunity, share of demo requests that name you before they arrive. Everything else is a means, and means should be allowed to change monthly.

2. 2

   Put the operator in the room

   The person changing the bids and writing the copy should be on the call. Account management exists to protect the agency's margin, not your outcome. If the operator is too senior to be on your account, you are paying for someone else's account.

3. 3

   Give explicit permission to cut

   Write into the engagement that reducing spend, killing a channel, or shipping less is an acceptable recommendation. Most agencies will not volunteer it. Almost all of them will act on it once it is safe.

The best argument for buying activity anyway

What good looks like from the client seat

You can feel the difference within a few weeks. Decisions get made in the thread instead of in the next monthly. Something you asked for gets pushed back on, with reasoning. The report gets shorter. Somebody tells you a test failed before you notice it yourself.

None of that is exotic. It is what an in-house team does, because an in-house team is not paid per deliverable. The agencies worth keeping are the ones that have decided to operate that way anyway and priced themselves accordingly.

The scope of work checklist

The onboarding question that predicts the whole relationship

Ask a prospective agency to walk you through their first thirty days on a new account. A vendor describes a kickoff deck, a brand questionnaire, and a media plan ready to launch by day fifteen. An operator describes a week spent reading your CRM, listening to sales calls, and finding the two or three things that are already working before touching the budget. The difference in that answer alone tells you more than three reference calls.

The vendor's plan is not wrong, exactly. It is just optimised for a different goal: getting spend live fast enough that the retainer looks justified in month one. The operator's plan looks slower on paper and is usually faster in practice, because the campaigns that launch in week three are built on what the account actually needs rather than a template built for the average client.

Pricing models and what they incentivise

Percentage of media spend is worth naming specifically, because it is common and rarely discussed openly. An agency paid 15 percent of ad spend has a direct financial interest in a bigger budget next quarter, independent of whether the current budget is producing pipeline. That is not a reason to avoid the model. It is a reason to ask the question directly: does your fee change if we spend less next quarter, and if the answer is a nervous pause, you have your answer.

What changes once the scope names an outcome

The clearest sign an engagement has switched from activity to outcome is what happens in the first monthly call after a bad month. In an activity contract, the call is a status update: here is what we shipped, here is why the metrics moved the way they did, see you next month. In an outcome contract, the call is a working session: here is what we think is broken, here is what we want to try instead, here is what we need from you to try it.

That second kind of call is uncomfortable in a useful way. It requires both sides to admit something did not work, which an activity contract never requires, because the activity happened regardless of the result. If your monthly calls have never once required anyone to admit something did not work, the contract behind them is still built on outputs, whatever the pitch deck said at the start.

A short note on switching agencies mid contract

If you recognise your current agency in the activity column above, the fix is rarely an immediate termination. Most contracts have a notice period, and most switches lose two to three months of momentum during the handover regardless of how the old relationship ended. The faster move is to renegotiate the existing scope toward an outcome before the contract is up for renewal, using the language in this post as the ask. Agencies that refuse to discuss outcome based accountability at all are telling you what the renewal conversation will look like too.

A second worked example, with the arithmetic

The fifteen thousand euro retainer example above shows the behaviour difference. Here is the number that usually settles the argument with finance: the fully loaded cost of the mistake, not just the invoice.

How to run this test on your own agency relationship

1. 1

   Pull the last two quarters of channel level pipeline data

   Not leads, not impressions. Opportunities created and their source, even if the attribution is imperfect. Imperfect and directional beats absent.

2. 2

   Calculate cost per opportunity by channel

   Rank every channel from lowest to highest. The bottom of that list is where the activity contract is protecting something that should have been cut two reviews ago.

3. 3

   Ask your agency to explain the bottom two channels unprompted

   Do not lead the witness. If the explanation is a plan to keep testing rather than a plan to reallocate, you are hearing the activity contract talk, not the operator.

4. 4

   Bring the two quarter cost number, not the monthly one

   A single month of waste is easy to shrug off. Two quarters, compounded against what the same budget could have produced elsewhere, is not.

The objection that comes up every time this argument is made