# Choosing a SaaS Marketing Agency

> A three week selection process with scoring weights, the eight questions that expose a weak shop, reference call scripts and how to verify case study numbers.

Source: https://saas-marketing.net/guides/how-to-choose-a-saas-marketing-agency/
Topic: SaaS Marketing Agencies
Type: guide
Published: 2026-09-11
Last updated: 2026-09-11
Publisher: SaaS Marketing (saas-marketing.net)
License: CC BY 4.0. Quote or republish with attribution and a link to https://saas-marketing.net/guides/how-to-choose-a-saas-marketing-agency/

## Short answer

Choose a SaaS marketing agency in three weeks: week one defines the single constraint you are buying against and sources five candidates, week two runs scored pitches against a twelve criterion rubric, week three runs reference calls and awards a paid pilot of five to ten thousand dollars. Verify every case study by asking for the baseline, the time window, the spend and who else touched the funnel. An agency that refuses a paid pilot is hiding a resourcing problem.

## Key takeaways

- Name one constraint before you shortlist: traffic, conversion, pipeline, or execution capacity. Agencies optimised for each are different companies.
- A paid pilot of 5 to 10 thousand dollars tells you more in six weeks than any RFP tells you in six.
- Case study numbers are meaningless without the baseline, the window, the spend and the list of other teams touching the funnel.
- Ask which account they lost last year and why. The answer separates shops that learn from shops that churn.
- Insist on meeting the people who will do the work, and check their names against the contract before signing.
- Twelve criteria, weighted, scored independently by three people before discussion, beats a debrief where the loudest voice wins.

---

Most agency selections are decided in the first fifteen minutes of the second pitch, by a founder who liked someone. Everything after that is procurement theatre built to justify the feeling.

That is not a criticism of instinct, which is often right. It is a criticism of a process that cannot tell you when instinct is wrong, and cannot survive the moment nine months later when results have stalled and nobody can remember what you agreed success would look like.

## Define the constraint before you talk to anyone

Week one, day one. Write one sentence naming the single thing that is blocking growth right now. Not three things.

There are really four constraints agencies get hired against, and they map to different firms.

The diagnostic is straightforward. Pull your last two quarters: sessions to key pages, conversion rate on those pages, meetings booked, and opportunities created. Whichever step has the worst ratio against comparable companies is your constraint. If every step looks bad, the constraint is execution capacity and you need hands before strategy.

A brief that says "grow pipeline" gets you five proposals that each describe a different job, and you will end up comparing a content calendar against a paid media plan against a website rebuild. Write the constraint, the baseline number, the target number, and the date. Everything a candidate proposes should be legible against those four facts. The [agency brief template](/templates/saas-agency-brief/) has the structure.

## Sourcing five candidates, not twenty

Five is the right number. Twenty means you are researching rather than deciding, and every additional candidate adds three hours of process for a diminishing chance of a better outcome.

Get them from three places. Referrals from operators at companies with your motion and ACV, which is the highest yield source by some distance, and worth asking for by name rather than posting in a Slack community. Competitor teardowns: run a competitor's domain through Ahrefs or Semrush, find who has been publishing or building links for them, and check the agency's own client list. And published work, meaning agencies whose people write publicly about the specific problem you have.

Discount directory listings heavily. Most "best SaaS marketing agency" lists are written by agencies that rank themselves first, which is why our own [best B2B SaaS marketing agencies](/guides/best-b2b-saas-marketing-agencies/) roundup discloses its criteria. If you are buying paid specifically, [SaaS PPC agencies](/guides/saas-ppc-agencies/) narrows the field further.

## The twelve criterion scorecard

Week two. Each candidate gets 60 minutes, structured, with the same brief and the same questions. Then three people from your side score independently before anyone discusses it.

| Criterion | Weight | What a strong answer looks like |
| --- | --- | --- |
| Motion fit (PLG, sales led, hybrid) | 15% | Names three clients with your motion and describes how the work differed |
| ACV band experience | 10% | Understands your sales cycle length without being told |
| Named delivery team | 15% | Introduces the actual strategist and writer, with current capacity disclosed |
| Diagnostic quality | 15% | Tells you something about your funnel you did not know |
| Measurement plan | 10% | Proposes a baseline and a month three checkpoint unprompted |
| Verified case study | 10% | Supplies baseline, window, spend and other active channels |
| What they would stop doing | 5% | Has a real answer about killing work that fails |
| Required input from you | 5% | Quantifies your team's weekly hours honestly, usually 4 to 8 |
| Pricing transparency | 5% | Clear scope, clear overage, no vague strategy line item |
| Contract terms | 5% | 30 to 60 day exit, IP assigned to you, no automatic annual renewal |
| Reference willingness | 3% | Offers three references including one that ended |
| Pilot willingness | 2% | Says yes immediately |

The weights matter less than the independence. Scoring before discussion stops the debrief becoming a negotiation between the loudest person and the person who found the agency. Use the [agency evaluation scorecard](/templates/saas-agency-rfp-scorecard/) rather than rebuilding it, and if you are running a formal process for procurement reasons, the [SaaS agency RFP process playbook](/playbooks/saas-agency-rfp-process/) covers the paperwork layer.

Notice that diagnostic quality carries the same weight as team. An agency that arrives having actually looked at your site and found a real problem has demonstrated more in ten minutes than a deck of logos demonstrates in an hour.

## The eight questions that expose a weak shop

Ask these in the pitch, in this order, and watch for hesitation rather than listening for content.

**Ask every candidate**

Question two is the one that does the work. A strong agency answers it in about four seconds, names the client, and gives a real reason: the champion left, the budget was cut, we were wrong about the channel, the relationship never had a clear owner on their side. A weak agency says they have never lost a client, which at any meaningful size is not true, or gives a version where the client was entirely at fault.

Question one exposes the bait and switch. The senior strategist in the pitch is often the founder, and the person who shows up in week three is a coordinator two years out of university. That is not automatically disqualifying, since good agencies influence juniors under real supervision. It is disqualifying when nobody will name them in advance or put them in the contract.

**8 questions** The pitch questions that separate delivery capability from business development polish

## Reference calls that produce information

Week three. Three references each for your top two, and you request one that ended, which is the call worth having.

Open with a question that gives permission to be honest: "I am not trying to get you to sell them to me, I am trying to find out where they are weak so I can manage it." Then ask what the first ninety days were like, what the agency got wrong, how they handled it, who from the agency was actually present by month six, and whether reporting ever changed shape after a bad month.

Two questions surface churn risk better than the rest. Ask which account the agency lost last year, and then ask the reference whether they would have hired them again knowing what they know now. A reference who pauses before answering the second has told you something.

One more, always: ask what the reference's own team had to do to make it work. Agencies fail as often from client side neglect as from their own capability, and if every reference says the agency needed six hours a week of subject matter expert time and you have none available, you have just learned the real risk in your own plan.

The full question bank sits in [agency vetting questions](/checklists/saas-agency-vetting-questions/), which is worth printing rather than improvising from.

## Verifying case study numbers

Four questions kill most inflated claims: what was the baseline, over what window, what was the spend behind it, and which other teams or channels were running at the same time.

"We grew organic traffic 400 percent" usually means a site went from 1,200 to 6,000 monthly sessions, over 18 months, while the company also raised a Series A, hired two in house writers, launched a free tool and got covered in a newsletter with 90,000 subscribers. All of that is fine. None of it is a reason not to hire them. It is a reason to understand what they contributed and what your own conditions need to be.

Ask for the numbers in absolute terms, not percentages. Percentages off a small base are the oldest trick in agency marketing and they are usually deployed innocently, because the deck was built by someone who was told to make it look good.

Then run the claim yourself. If they say they grew a client's organic traffic, put that client's domain into Ahrefs or Semrush and look at the curve and the dates. Two minutes, and it either confirms the story or ends the conversation.

## Why a paid pilot beats everything else

Six to eight weeks, five to ten thousand dollars, one defined deliverable that has standalone value.

Good pilots: a full keyword and content gap analysis with a prioritised twelve month plan, four published articles through your own review process, a conversion audit of your five highest traffic pages with implemented changes, or a paid campaign build with the first four weeks of spend managed.

The pilot tests the things a pitch cannot. How fast do they reply on a Thursday afternoon. Does the strategist from the pitch still attend calls in week five. Is the first draft usable or does it need a rewrite that costs you more time than writing it yourself would have. How do they take critical feedback, which is the single most predictive signal in the whole process.

An agency that refuses a paid pilot will explain that their process needs a longer runway to show results. Sometimes that is true, particularly for SEO. It is still not a reason to refuse, because the pilot deliverable is the plan, not the result. Refusal almost always means they are fully staffed and the pilot would cost them a body they do not have, which is exactly the resourcing problem you want to know about before signing twelve months.

The obvious tradeoff: pilots cost you six to eight weeks of calendar time, and if you are three months from a funding milestone you may not have it. In that case shorten the pilot to a single deliverable in three weeks rather than skipping it. Also budget for the possibility that the pilot fails and you restart with candidate two, which is a good outcome that feels like a bad one.

## The build versus buy check you should run first

Before any of this, run the numbers on doing it in house. An agency at 15 thousand a month is 180 thousand a year, which buys a senior content marketer plus a freelance budget, or a demand gen manager plus tooling.

Agencies win on speed to start, breadth of skill, and the ability to stop. In house wins on product knowledge, compounding institutional memory, and cost at steady state. The crossover usually arrives when the work becomes predictable, which for content is around month twelve. Run your own figures through the [agency vs in house cost calculator](/calculators/agency-vs-in-house-cost/) rather than arguing about it abstractly, and read [what a B2B SaaS marketing agency actually does](/guides/b2b-saas-marketing-agency-agencies/) if the scope boundaries are still fuzzy.

## The three week calendar

Week one: write the constraint and the brief, source five candidates, send the brief with a deadline. Week two: five scored 60 minute pitches, independent scoring, shortlist two. Week three: three reference calls each, verify case studies, award the pilot.

Then six to eight weeks of pilot, one honest review meeting, and only then an annual contract with a 30 to 60 day exit clause. If that feels slow, compare it to the twelve months you lose when the wrong agency takes until month nine to become obviously wrong. Start from the [SaaS marketing agencies overview](/saas-marketing-agencies/) if you are still deciding whether to hire at all.

## Frequently asked questions

### How long should choosing a SaaS marketing agency take?

Three weeks of focused work, plus a six to eight week paid pilot before any annual commitment. Selections that run three months usually do so because nobody defined the constraint at the start, so the criteria kept shifting. Compressing to one week is the opposite failure: you end up choosing on chemistry because there was no time to run references.

### What should I ask a SaaS marketing agency before hiring them?

Ask who specifically will work on the account and what else they are staffed on, which client they lost last year and why, how they would measure success in month three, what they would need from your team weekly, what they would stop doing if results stalled at month six, and to walk through one case study with baseline, window and spend disclosed. Vague answers to any of these are the signal.

### How much does a SaaS marketing agency cost?

Retainers for B2B SaaS specialist agencies commonly run 8,000 to 30,000 dollars a month, with content focused shops at the lower end and full funnel or demand generation agencies at the upper. Paid media agencies often charge a percentage of spend, typically 10 to 15 percent, with a floor. A paid pilot should cost 5,000 to 10,000 and produce a defined deliverable.

### Should I run an RFP or a paid pilot?

A paid pilot, in almost every case. An RFP measures how well an agency writes proposals, which is a skill held by the business development team rather than the delivery team. A pilot measures the actual working relationship: responsiveness, quality of the first deliverable, and whether the senior person on the pitch is still present in week four.

### How do I verify an agency case study?

Ask four questions: what was the baseline before you started, over what time window did the change occur, what was the media spend or content volume behind it, and which other teams or channels were active at the same time. Then ask to speak to that client. Agencies with real results answer all four quickly. Agencies without one start explaining why the numbers are confidential.

### Does a SaaS marketing agency need experience in my exact category?

Category experience matters less than motion experience. An agency that has run product led growth for three developer tools will outperform one that knows your vertical but has only worked with sales led enterprise vendors. Ask about ACV band, sales motion and buyer persona rather than industry label. The exception is regulated categories, where compliance review cycles genuinely change the work.