Best direct marketing agencies compared, and the one number that has to be agreed before anybody writes a brief
Direct marketing is the discipline where every campaign is judged on measured response, so the supplier conversation is really an argument about what counts as a response and who gets credit for it.
Published by AI Knows Us · Updated 22 October 2026
A direct marketing agency sells accountable response across channels rather than brand work: the campaign is designed so that a specific action can be counted, attributed and priced. The decision that matters when hiring one is not the agency's creative reel. It is whether you and they have agreed, in writing and before the brief, what a qualified response is and what you are willing to pay for one. Get that wrong and every report afterwards is an argument.
This page covers what sits inside the phrase, the three kinds of shop that answer the enquiry, how the fee models change the agency's behaviour, and the questions that sort a shortlist. If your campaign is specifically physical mail, the supplier types and unit economics for that channel are different and are covered in direct mail companies and the cost structure that decides which one to use. This page is the broader discipline.
What direct marketing actually covers now
The phrase predates digital advertising and it has survived because it describes a method rather than a channel. A direct marketing campaign has three properties: it addresses an identified person or a tightly defined audience, it asks for a specific action, and the action is counted. That covers addressable mail, email to an owned list, outbound telephone, direct response television and radio, inserts, SMS, and addressable digital where the targeting is built from your own data rather than bought as reach.
What it does not cover is brand advertising, where the measurement is lift rather than response. Agencies are honest about this when asked directly, and the mismatch is the most common cause of a bad first quarter: a brand shop given a cost per acquisition target, or a response shop asked to build fame.
Three kinds of shop, and what each is actually good at
| Kind | Where its skill is concentrated | Right for | Where it is weak |
|---|---|---|---|
| Classic direct response agency | Offer construction, creative that converts, disciplined testing, control versus challenger | A proven product with a repeatable offer, where the gain comes from beating the current control | Audience and data engineering, which it often subcontracts |
| Data and CRM led agency | Audience selection, segmentation, modelling, suppression, lifecycle triggers | A large customer file where the gain is in who you contact rather than what you say | Creative range. The work is often correct and unmemorable |
| Performance and channel specialist | Buying and optimising inside platforms day to day | Scaling something that already converts, fast | Offer and proposition work, which it tends to treat as given |
Most briefs need two of the three, and the sequence matters: the audience and the offer have to be settled before any media is bought. An agency that wants to start with channel allocation is answering a question you have not asked yet.
The fee model, because it decides what the agency optimises for
- Monthly retainer for a defined scope. Predictable and neutral. It rewards nothing in particular, which is usually what you want while you are still learning what works.
- Percentage of media or volume. Simple, and it quietly rewards spending more and mailing more. Acceptable with a volume cap and a cost per response ceiling written in.
- Project fee per campaign or per test. The cleanest way to buy a first engagement, and it makes a second one easy to justify or refuse.
- Cost per response or per acquisition. The model buyers ask for and the one that needs the most care. It only works when the definition of a response is unambiguous, the attribution window is agreed, and duplicate and unqualified responses are excluded by a rule rather than by argument. Without those three, you will spend the quarter disputing invoices.
How a fee model changes agency behaviour is not unique to this discipline, and the pattern is worth reading in full in the fee model decides how the agency behaves.
The criterion to rank agencies on, applied to all of them
The checkable criterion in this discipline is whether the agency will publish, before the contract, the test design it intends to use to prove the campaign caused the response. Not the creative approach. The test. A holdout group, a matched region, a scheduled on and off period. A direct marketing agency that cannot describe how it will separate its own effect from everything else in the market is selling creative services with a response vocabulary.
Ranked on that criterion, the data and CRM led shops tend to come first, because holdouts are native to their craft, and the channel specialists often come last, because platform reported conversions are the easiest number to hand over and the hardest to defend. The arithmetic for turning a measured response into a figure a finance team accepts is set out in the two opposite searches hiding in one lead generation phrase.
Six questions that sort the shortlist
- What is a response, in one sentence, and what is excluded? Write both answers down and attach them to the contract. Everything else follows from this.
- What is the attribution window, and who owns the tracking? A thirty day window and a seven day window describe two different campaigns at two different prices.
- How will we know the campaign caused it? Ask for the holdout design, and ask what it costs in forgone volume. A real answer includes a cost.
- Who builds the audience, and whose data is it afterwards? Suppression files, models and segment definitions should come back to you on exit, named in the contract.
- Show me a test that failed and what you concluded. Direct marketing is the one discipline where failed tests are the asset. An agency with no failures has not been testing.
- What is the smallest first engagement you would take? A supplier who insists on a twelve month retainer for a first test is pricing its own risk into your learning.
The response channel that has no attribution model yet
Direct marketing is built on a simple premise: you can see what happened because you asked for an action and counted it. There is now a step before the action that nobody is counting. A person who sees your mail piece, your DRTV spot or your email does not always respond to it directly. A growing share of them go and ask an assistant about you first, and what ChatGPT, Gemini, Perplexity, Claude or Google's AI Overviews say back is not controlled by the campaign and is not in anybody's response report. An offer that works, answered by a sentence that undersells you, shows up in the data as a weak creative test.
It is checkable in the same spirit as everything else in this discipline: ask the questions a prompted buyer would ask, record which businesses get named and which sources the answers lean on, and take the reading before the campaign so you have a baseline. AI Knows Us reads those five surfaces on a schedule and reports the sources behind each answer, and the free scan on your own domain needs no account. On the published plans, Growth at 299 dollars a month watches 100 buyer questions across all five surfaces, and a Claude check uses 3 credits because Claude costs more to run. Where a rival beats us for a direct marketing team specifically: Profound publishes "SSO/SAML + SOC 2 compliance" on its enterprise tier and tracks "Up to 9 Answer Engines" against our five, read on tryprofound.com/pricing on 21 October 2026. If your data protection review requires SSO or a SOC 2 report, our rendered pricing text carries neither, and theirs is the right purchase.
Questions buyers ask about direct marketing agencies
What is the difference between a direct marketing agency and a performance agency?
Direct marketing starts from an audience and an offer and picks channels afterwards. A performance agency starts inside a channel and optimises it. Both count responses, and only one of them will argue with you about the offer.
Is direct marketing still worth it when email is free to send?
Sending is cheap, which is exactly why list quality, suppression and offer discipline decide the result. The cost moved from postage to attention, and the agencies that are good at this are the ones that treat a send as expensive anyway.
Should we pay on a cost per acquisition basis?
Only once you have run a few campaigns and know your own numbers. On a first engagement the agency will price its uncertainty into the rate, so you pay a premium for a model you cannot yet verify.
How big does our list need to be for a test to mean anything?
Big enough that a difference between two cells is larger than the noise, which depends on your response rate rather than on a universal number. Ask the agency to state the minimum cell size for the response rate it expects, before the test.
Who should own the data in a direct marketing relationship?
You, in every case, including suppression lists and any model the agency builds from your file. Agree it before the first campaign, because it is an awkward conversation to have during an exit.