Growth · 11 min read

How to Evaluate B2B Marketing Companies

A vendor evaluation scorecard for B2B marketing companies: the five types, how to read a proposal, the red flags that predict failure, and the questions to ask on the first call.

The short answer

Evaluating B2B marketing companies fails when you compare across types instead of within one. Sort candidates into the five types first, shortlist three inside the type that fixes your constraint, then score every one on the same five dimensions: constraint fit, named ownership, a dated 90-day deliverable list, whether reporting ends at pipeline, and whether the out-of-scope list is written down. Choose on the out-of-scope list, because that is what you manage around for six months.

  • Compare within a type, never across types, or the proposals cannot be scored against each other
  • A dated 90-day deliverable list is the single most predictive artifact in any proposal
  • The out-of-scope list matters more than the capability list, because it is what you live with
  • Reporting that ends at impressions and activity counts is the most common structural failure

Free vendor kit

The B2B growth vendor kit

Two files we use with real prospects: the raw benchmark dataset behind our AI search readiness research, and the RFP template we hand to teams running a paid media search. One email unlocks both.

  • AI Search Readiness Benchmark 2026 dataset CSVPer-domain scores for 160 live software sites across 10 readiness signals, fielded 2026-08-09. Free to cite under CC BY 4.0 with a link.
  • B2B paid advertising RFP template PDFEight pages: scope sections, the scoring rubric with weights, the questions that separate operators from resellers, and the automatic disqualifiers.

Where should we send the kit?

Work email unlocks both files on the next page. We use it to send the next version of the research, nothing else.

No newsletter drip, no reselling your details. Both files download on the next page.

Vendor evaluation in B2B marketing usually goes wrong in the first week, before anyone has read a proposal. A team collects five referrals, books five calls, and ends up holding five documents that describe five different jobs at five different prices. There is no way to score that set, so the decision falls to whoever presented best.

The fix is structural, not diligence effort. Sort by type, shortlist within a single type, and score everyone on the same five dimensions. This guide walks that process, including the artifacts to demand and the specific answers that predict a bad engagement. We run one of these types ourselves, an integrated firm, so treat the positioning here as informed and biased. The full type comparison lives on our B2B marketing companies page.

Step one: name the constraint, not the goal

Every team's goal is more pipeline. That is not useful for evaluation because every firm sells it. The constraint is the one thing that, left unchanged, prevents everything else from working. There are four common ones in software:

  • Nobody owns the whole picture. SEO, content, paid, and web sit with four vendors. Each hits their number and pipeline is flat.
  • The site cannot support the program. Content is fine and rankings are capped anyway: slow pages, no internal link architecture, a CMS that needs a developer for a headline.
  • The positioning is unclear. Traffic converts poorly everywhere, sales calls start with confusion, and no channel fixes that.
  • Capacity. The plan is right and there is one of you.

Each constraint maps to a different type of company. Buying an excellent fix for the wrong constraint is the single most expensive mistake in this category, and it is nearly always a sorting error rather than a judgment error.

Step two: shortlist inside one type

Three candidates from one type produce a real comparison. One candidate from each of three types produces a preference. If your constraint is fragmentation, talk to three integrated firms. If your constraint is positioning, talk to three product marketing consultancies and revisit channels afterward.

This also fixes the price confusion. Within a type, ranges cluster tightly, and an outlier means something specific: unusual seniority, unusual scope, or a scope gap you have not found yet. Across types, price tells you almost nothing. Our own breakdown of agency pricing covers what the ranges buy at each level.

Step three: the five-dimension scorecard

Score each candidate one to five on the same five dimensions. Total the scores, but pay more attention to which candidate you cannot score, because an unscoreable candidate is selling a deck rather than a delivery model.

  • Constraint fit. Does this firm's core build address the sentence you wrote in step one, or does it address it as an adjacent capability.
  • Named ownership. Who does the work after signature, by name and role. A strategist who appears only in the pitch is a sales asset.
  • Dated 90-day list. Specific deliverables with dates. A phase diagram labeled discovery, strategy, execution is not a plan.
  • Measurement depth. Does reporting terminate at pipeline influenced or at impressions, rankings, and activity counts.
  • Scope honesty. Is there a written list of what they will not do. Firms that refuse to write one negotiate it later, at your expense.

How to read a proposal

Proposals in this category are optimized to look thorough. Read them in a fixed order and most of the volume becomes skippable.

  1. Find the deliverable list. If there are no dates and no counts, stop reading and ask for them before the next call.
  2. Find the team page and check whether roles are named or generic. "Senior SEO strategist" without a name means unassigned.
  3. Find the metric the engagement is judged on. There should be exactly one primary, with leading indicators underneath it.
  4. Find the exclusions. Media spend, licenses, and third-party fees should be explicitly outside the retainer at any honest firm.
  5. Find the term and the exit. A six-month minimum is reasonable for compounding work. An auto-renewing twelve-month term with a ninety-day notice window is not.

The questions with a cost to answering

Capability questions are rehearsed and every firm passes them. Use questions where an honest answer costs the firm something.

  • What would you cut from our current program first, and what evidence triggers it.
  • Describe an engagement that failed and what you changed structurally afterward.
  • Which prospective clients do you turn down, and why.
  • If every leading indicator is flat at day 90, what specifically happens.
  • Who on your team will I speak to in month four, and how often.

The failure question is the most informative one in the set. A firm with no failure story either has not run enough engagements or is unwilling to be accurate with you, and both answers matter.

Red flags that reliably predict a bad engagement

  • Pricing withheld until a qualification call. The number is being set from your funding announcement rather than your scope.
  • Percentages without baselines. A 340 percent traffic increase from a base of 900 sessions is a rounding event.
  • Guarantees on organic outcomes. Nobody controls a ranking or a citation. Guaranteed deliverables are fine; guaranteed results are not.
  • Pitch team disappears at signature. Ask directly, then confirm the answer in the contract.
  • No out-of-scope list. Every scope has edges. A firm unwilling to write them down is preserving room to renegotiate.
  • Channel answers with no tradeoffs. A firm that recommends everything has not diagnosed anything.

A worked example of the scoring

A Series A software company with four vendors and flat pipeline writes its constraint as fragmentation. It shortlists three integrated firms at $9K, $14K, and $22K per month.

The $9K firm scores well on constraint fit and price, has one named strategist across eleven accounts, and its 90-day list contains four content pieces and an audit. The $22K firm names five people, delivers a dated list with a site rebuild inside the first 60 days, and reports to pipeline influenced. The $14K firm cannot produce a dated list at all and describes measurement as "full-funnel visibility".

The middle option drops out on unscoreability, not price. The real decision is between capacity risk at $9K and budget risk at $22K, and it turns on whether the site itself is part of the constraint. If it is, the cheaper option cannot fix the thing that is broken and the price difference is irrelevant.

Reference checks that produce signal

Ask for two references: one current client past the two-year mark and one client who left. The second request is the test. A firm that cannot produce a former client who will speak either has never retained anyone long enough to churn well or does not want you talking to the ones who did.

On the call, skip outcomes and ask about mechanics: what the reporting cadence was in month four, what got missed and how it was handled, whether the seller stayed involved, and what they would scope differently. Those answers describe delivery. Outcome anecdotes describe the sale.

What a good first 90 days looks like

Regardless of type, a competent engagement front-loads diagnosis and instrumentation so spend follows measurement rather than preceding it. Month one should produce working analytics and CRM handoff, a mapped query and channel set, and a technical and AI-readiness baseline. Months two and three should produce shipped pages, published content on a stated cadence, and live campaigns with weekly iteration.

By day 90 you should be able to check leading indicators agreed before signature: pages shipped against plan, index and citation coverage, ranking movement on the mapped set, conversion rate on money pages, and cost per qualified conversation. If all of them are flat, the lag argument stops protecting the program. Our month-by-month methodology shows the version of this we commit to in writing.

Making the call

Pick the candidate whose out-of-scope list you can live with. Capability lists converge across serious firms; exclusion lists are where the real differences sit, and they are what you manage around every month for the length of the term.

If your constraint is fragmentation and you sell software, that is the case we are built for: website, SEO, AEO, GEO, content, and paid under one strategist with published pricing and a written scope. Bring your current vendor mix and last quarter's reporting to a 30-minute growth call and we will tell you which type of company you actually need, including the cases where it is not us. If you would rather read first, the agency archetype guide covers the diagnosis step in more depth, and what best actually means for your company explains why ranked agency lists cannot answer it.

Ready to run this playbook?

Momentence runs the full engine, one team, one dashboard, six-month minimum. 30-minute call, no pitch deck.

Free vendor kit

The B2B growth vendor kit

Two files we use with real prospects: the raw benchmark dataset behind our AI search readiness research, and the RFP template we hand to teams running a paid media search. One email unlocks both.

  • AI Search Readiness Benchmark 2026 dataset CSVPer-domain scores for 160 live software sites across 10 readiness signals, fielded 2026-08-09. Free to cite under CC BY 4.0 with a link.
  • B2B paid advertising RFP template PDFEight pages: scope sections, the scoring rubric with weights, the questions that separate operators from resellers, and the automatic disqualifiers.

Where should we send the kit?

Work email unlocks both files on the next page. We use it to send the next version of the research, nothing else.

No newsletter drip, no reselling your details. Both files download on the next page.

FAQ

Common questions.

Write your constraint in one sentence, decide which of the five company types fixes it, then talk to three firms inside that type only. Score each on constraint fit, named ownership, a dated 90-day deliverable list, whether reporting ends at pipeline, and whether the out-of-scope list is written. Three comparable conversations beat six incomparable ones and take about three weeks.

A named strategist and named executors with roles, a dated deliverable list for the first 90 days, the one metric the engagement is judged on, leading indicators to check at day 90, a written out-of-scope list, the pricing with any one-time fees stated, and what happens if leading indicators are flat. Anything less is a capability brochure.

Pricing that only appears after a qualification call, percentage-based case studies with no baselines, guaranteed rankings or guaranteed pipeline inside 90 days, a pitch team you will never see again after signature, no answer to what they will not do, and no account of an engagement that failed. Any two of these together is usually enough to drop a candidate.

Ask for a client who left and a client who stayed past two years. Ask the reference what the firm got wrong and how it was handled, what the reporting cadence actually was after month three, and whether the person who sold the engagement remained involved. References volunteering only outcomes are describing the sale, not the delivery.

Six months minimum for anything involving organic search, content, AI search visibility, or a site rebuild, because the work front-loads and the return back-loads. A three-month term usually means you pay for the build and leave before it produces anything. Paid-only engagements can start shorter because the feedback loop runs in weeks.

Keep reading

More from Learn.

Search · 18 min read

The Definitive SEO, AEO, and GEO Guide for Software

One operating system for SEO, AEO, and GEO, with actionable checklists for technical foundations, answer coverage, entity data, off-page corpus work, and measurement.

Paid · 11 min read

B2B PPC Agency Pricing: Fee Models and Real Cost

How B2B PPC and paid media agencies charge, what each fee model rewards, worked total cost math at $15K, $40K, and $100K monthly spend, and the break-even point against an in-house buyer.

Growth · 11 min read

What Makes the Best B2B Marketing Agency for You

Best is situational, not a ranking. The five decision axes that separate B2B marketing agencies, a scoring rubric, disqualifiers, and how to read the shortlist you already have.

Growth · 12 min read

B2B Marketing Firm vs In-House Team: How to Decide

A build-versus-buy framework for B2B software companies: real total cost of ownership, a scoring model, the hybrid setups that work, and how to hand execution back in-house later.

Growth · 12 min read

How to Choose a B2B Marketing Agency

A diligence framework for choosing a B2B marketing agency: the four agency archetypes, what each fixes, the questions that expose weak shops, and the real cost of fragmentation.

Buying Guide · 12 min read

SaaS Marketing Agency Pricing: What Each Retainer Buys

What SaaS marketing agency pricing actually covers at every retainer band, how agencies build a price, and how to audit a proposal before you sign it.

Paid Media · 12 min read

How to Choose a B2B SaaS Paid Advertising Agency

How to evaluate a B2B SaaS paid advertising agency: fee models, minimum viable spend, measurement prerequisites, and the diligence questions that matter.

AEO · 10 min read

GEO vs AEO: Two Different Jobs Inside AI Search

GEO and AEO are related but not the same. AEO wins the citation on a specific answer, GEO shapes how models describe your brand. How to run both together.

Demand Gen · 8 min read

Inbound vs Outbound for B2B SaaS: How to Split the Budget

A practical framework for splitting growth budget between inbound demand capture and outbound demand creation, with the failure modes of each.

SEO · 10 min read

Best SaaS SEO Agencies in 2026

A criteria-driven comparison of the SaaS SEO agencies worth shortlisting in 2026: where each is strongest, where it is not, and how to pick one.

AEO · 9 min read

AEO vs SEO: The Difference and Why You Need Both

AEO is the discipline of getting cited by ChatGPT, Perplexity, Gemini, and AI Overviews. How it differs from SEO and how to run both together.

Website · 10 min read

How to Build an AI-Native Marketing Website

Ship a modern software marketing website in weeks: AI in the production loop, SSR performance from day one, and a design system marketing can extend.

Demand Gen · 11 min read

Software Demand Generation Playbook for 2026

How B2B and B2C software companies combine site, SEO, AEO, content, and paid into one demand gen system that compounds instead of resetting.