Direct mail agency vs. DIY platform: Which is right for you?

By Jeff Tarran, Integrator, Gundir

You’ve decided direct mail is worth testing. Now you’re facing a decision that trips up more marketers than it should: do you hire a full-service agency and let them own the program, or sign up for a self-serve platform and run it in-house?

And if you pick wrong, are you looking at a small inefficiency, or a program that underperforms for months without understanding why?

As a strategist for Gundir, I’ve run both tracks. We build and manage full-service direct mail acquisition programs for companies in financial services, healthcare, SaaS, and subscription businesses. We also offer GundirXpress™, our own self-serve platform, for marketers who need professional production at a lighter scale.

I’ve watched this decision play out often enough to know the answer almost always comes down to four things:

1. your volume
2. your in-house capability
3. your budget
4. your timeline

Here’s exactly how to use them.

What a full-service direct mail agency provides

A full-service agency owns the entire program (strategy, creative, data, production, and analysis) so you don’t have to build each of those capabilities yourself. That sounds like a broad claim until you see how many distinct disciplines a functioning direct mail program requires.

A typical engagement runs through six areas:

1. goal-setting and KPI definition
2. audience strategy and test design
3. data modeling and list acquisition
4. direct-response creative development
5. production and print management
6. ongoing program management

Each area requires skills most marketing teams don’t have in-house. It’s rarely  the case that a successful digital marketer can transfer success to the DM channel without help. 

When one of the above  areas  is weak (and this is where programs fail most often), the rest of the investment doesn’t produce results it should.

In my experience, these are the three most common failure point for DIY direct mail: 

1. Data — For simple data targeting, such as geo or age, the data options available on DIY platforms are likely to be fine. But, many businesses can benefit by mailing to better defined data sets.
   
   Physical address-based data is the single best source of individual contact information. Because of that, the wealth of data that can be appended to any one mail record far exceeds what is available to digital records.
   
   For many products, this ability to laser target is the difference between mediocre and superior results.

Businesses with existing customer files can also benefit from modeling: using their customer file as the basis for developing a mail file that most resembles their existing customer base. In the majority of cases, a look-a-like model of this nature will outperform any off-the-shelf data set.

2) creative.

Companies assume their digital design team can adapt existing brand assets for mail. Digital designers optimize for aesthetics; direct-response designers optimize for action:

• outer envelope strategy
• offer hierarchy
• response mechanism placement (CTA)

These aren’t aesthetic preferences, they’re the conversion levers that determine whether a mail piece gets opened or recycled. (I’ve seen beautifully designed mail pieces with response rates that would make you wince.)

3) Strategy and Analytics

Did it work? Why or why not? What’s the next thing to test? DIY platforms are great for getting mail out the door. But, if learning and growing the channel are important goals, then test strategies and analytics are essential.

At Gundir, full-service programs are built around campaigns of 100,000 pieces or more. That’s the minimum volume for test cells to reach statistical significance. Below that threshold, you’re drawing conclusions from noise.

What a self-serve direct mail platform provides

A self-serve platform handles printing, postage, and tracking, and leaves strategy, creative, and analysis to you. That’s the right split for companies that have those skills in-house and need professional-grade execution at lower cost.

Gundir’s self-serve platform, GundirXpress™, starts under $1 per piece and includes templates, list tools, printing, postage, and campaign tracking. For smaller budgets, simpler campaigns, or companies not yet ready for full-service, it’s a practical starting point. (I’ll point you toward it even when the answer works against our full-service model. That’s how this should work.)

What a self-serve platform won’t do: design your test architecture, run look-alike data modeling, employ best-practices creative development, or explain why a campaign underperformed. Those are agency capabilities. If you need them and don’t have them in-house, a platform will get you into market faster but won’t help you learn from what happens when you arrive.

Agency vs. platform: how they compare

Here’s how the two tracks differ across the factors that matter most for this decision:

An agency provides the thinking and the execution. A platform provides the execution. If you have the thinking in-house, you don’t need to pay for something you already have.

When a full-service agency makes sense

A full-service agency is the right call when direct mail needs to become a real acquisition channel, not a one-time experiment. Four situations where this is clearly correct:

• Your campaign volume is 100,000 pieces or more. At that scale, the agency’s test discipline starts paying for itself. Below that volume, the math usually doesn’t favor the agency model.
• You don’t have direct mail expertise in-house. Data modeling, direct-response creative, production management, and attribution setup are specialized skills. Digital marketing experience doesn’t automatically transfer.
• You’re trying to beat a degrading control. Companies with direct mail programs that have plateaued need a structured beat-the-control process, not a fresh template. That work requires campaign history and experienced creative development.
• You’re acquiring cold prospects at scale. Top-of-funnel acquisition success via mail often depends on look-alike data modeling. We’ll test a number of different models and have found that model quality can increase response rates by  0.3% to   1.2%. That’s agency work.

“Mochi Health: 53% CAC reduction within one year. Full-service program — data, creative, test design, and ongoing optimization across multiple drops.”

Mochi Health needed a direct mail program that could drive subscription sign-ups at a CAC their unit economics could support. In calendar year 2025 , we reduced their cost per acquisition by 53%. That result came from the full program — data, creative, test design, and optimization across multiple drops — not from any single piece of it.

When a platform makes sense

A self-serve platform is the right choice when your volume, budget, or in-house capabilities make a full-service engagement more than you need right now. Three situations where I’d point you toward a platform first:

• You’re mailing fewer than 50,000 pieces. At that volume, test cells can’t produce statistically reliable results, so the agency’s test discipline doesn’t apply. Get in market, observe what happens, and revisit a full-service engagement when the scale justifies it.
• You have direct mail expertise in-house. If your team has direct mail creative experience, a working suppression file, and a clear attribution plan, a platform gives you professional production without paying for services you already have.
• Your timeline won’t accommodate 8–12 weeks. Full-service programs take time, with a first drop typically 8–10 weeks from signed contract, 11–12 weeks if data modeling is included. If your window is shorter, a platform can get you in market within days.

GundirXpress starts under $1 per piece as of June 2026. For experienced in-house teams that need professional production, or for companies testing the channel before committing to a full program, it’s the starting point.

The hybrid approach

Not every company fits cleanly into one track, and that’s fine.

A hybrid approach uses agency expertise for strategy and creative while relying on a platform for production execution. You get the specialized skills that are hard to replicate in-house (test architecture, look-alike modeling, direct-response copy), without necessarily committing to a fully managed program.

We see this with companies that have done enough direct mail to know what they don’t know, but aren’t ready to hand off the entire program. The model can work.

One thing to get right: attribution. If strategy and execution are split across relationships, whoever owns the measurement setup needs to own it from day one, not configure it retroactively after results come in. Matchback analysis, personalized QR codes, holdout cells: these decisions have to happen before the first drop.

Frequently asked questions

Q: Is a direct mail agency worth the cost for a first campaign?

A: It depends on what “first campaign” means. If it’s a structured test designed to produce a scalable program, an agency earns its fee; that’s exactly the work they’re built for. If it’s a 10,000-piece postcard to see whether direct mail does anything, start with a platform. The agency’s test discipline only pays off at volumes where test cells can reach statistical significance.

Q: Can I provide my own creative to a full-service agency?

A: Yes, and at Gundir we review client creative against direct-response principles before running it; the goal is mail that works, which means it needs to be engineered for response. In a number of cases, client creative has outperformed agency creative. It happens when clients have genuine DM experience in-house. What we won’t run is creative that isn’t built for action, regardless of who developed it.

Q: How long does it take to see results from a direct mail campaign?

A: From signed contract to first drop is typically 8–10 weeks with a full-service agency, 11–12 weeks if data modeling is included. After the drop, plan 30–60 days for matchback analysis to reconcile responses to mail records. A realistic first meaningful read is 12–16 weeks from project start.

Making the call

Up until 2022 , the most common question I heard was whether direct mail still worked. That question is largely settled. The category has shifted: teams that have decided to test the channel want to know how to run it well.

You now have the criteria to decide between working with an agency or running your own direct mail program:

1. volume
2. in-house capability
3. budget
4. timeline

If at least two of those four point toward an agency, that’s the right direction. If none of them do, start with a platform and revisit when the scale justifies it.

The next question most teams hit is how to structure the first test: which variables to isolate, how to set up attribution before the drop, and what a good result actually looks like. That’s where the real program-building starts.

How to take action now:

• If your planned volume is under 50,000 pieces, start with a self-serve platform before committing to an agency model
• If you’re at 100,000 pieces or more, have a strategy conversation with an agency before spending anything on production
• Decide who owns attribution upfront; matchback, QR codes, or holdout cells need to be configured before the first drop, not after
• If you’re evaluating both tracks, request a GundirLaunch™ consultation; the first conversation is about your program design, not a pitch

[INTERNAL-LINK: “how long does a direct mail campaign take” → gundir.com post on campaign timelines]

Sources and references

This article is based on Gundir’s direct experience managing direct mail acquisition programs across financial services, healthcare, SaaS, and subscription verticals. The Mochi Health result (53% CAC reduction within one year) is from Gundir’s published case studies. GundirXpress pricing (under $1 per piece; see TODO-5 above for date) and full-service timeline figures (8–12 weeks to first drop) are from Gundir’s documented service specifications.