How to Vet B2B Marketing Agencies for Pipeline Impact

How to Vet B2B Marketing Agencies for Pipeline and Revenue Impact

To separate agencies that drive pipeline from agencies that drive decks, run these 5 procedures in order. You will need CRM data definitions, 2 hours of executive time per finalist, and 3 to 5 reference clients. The full cycle takes 7 to 10 business days. The Starr Conspiracy recommends starting only after you have documented deal size, sales cycle length, and target vertical. See our demand generation glossary entry for shared definitions.

Step Summary

1. Audit case studies for industry, deal size, and cycle length relevance.
2. Extract CRM-backed attribution proof for pipeline and CAC payback.
3. Run structured reference calls with revenue-accountable buyers.
4. Read client retention signals as proxy performance data.
5. Stress-test channel-specific proof across SEO, content, ABM, and email.

Run all 5 on every finalist. Skip none. Negotiate none.

Most B2B marketing executives skip three of these five and then act surprised when 12 months of retainer produces a brand refresh and a pile of MQLs sales refuses to call. Run this before procurement locks the SOW. You are personally accountable to a CFO who will not accept "engagement" as proof and a board that scores you on pipeline. I have watched teams burn two quarters arguing about engagement metrics while pipeline flatlines. The procedures below are the standard we hold our own work to at The Starr Conspiracy, because we operate in the same complex verticals our clients do: SaaS, fintech, and manufacturing, with sales cycles that run 6 to 18 months and committees of 7 to 12 people. Vet us by this exact process. We expect it. This is how you verify an agency can build a marketing system, not run campaigns. We do not sell AI experiments. We build systems that prove pipeline in the CRM.

A quick aside on the categories of agency failure you are screening out. Luddites who refuse modern attribution and call it "creative integrity," Tourists who pitch every vertical and prove none, and Zealots who pitch AI as a shortcut around fundamentals. All three lose you a year.

Prerequisites and What You Need Before Starting

Before the first procedure, assemble the following. Skip any of these and the rest of the workflow produces noise.

• Your last 4 quarters of pipeline data, segmented by source, with definitions for SQL, opportunity, and closed-won.
• Written numbers for current CAC, CAC payback period, and average deal size by segment.
• A one-page profile of your ideal client: industry, revenue band, buying committee composition, average sales cycle.
• Authority to request anonymized, aggregated, permissioned CRM exports under NDA from finalist agencies. The agency's client must approve any reference call. Do not request or share personally identifiable data.
• 2 hours of CFO or RevOps time to validate any financial claims a finalist makes.
• A working understanding of demand generation versus brand marketing, because most agencies will conflate them.

If an agency refuses to sign an NDA that allows them to share anonymized client attribution data, the vetting process ends there. That refusal is the answer. Run this the moment you have 3 finalists.

Step 1, Audit Case Studies for Industry, Deal Size, and Cycle Length Relevance

Pull every case study the agency publishes. Filter on three variables before reading a single word of narrative: industry vertical, deal size, and sales cycle length. A case study from a $5K ACV self-serve SaaS does not predict performance for a $250K enterprise fintech sale. Different committees, different channels, different attribution windows.

This filter matters because the wrong-vertical agency will spend the first 6 months of your retainer learning your market on your dime. In fintech, that means regulatory review cycles they have never navigated. In manufacturing, procurement gates they have never broken through.

For each remaining case study, confirm three things. The named client is verifiable on LinkedIn and the named contact will take a reference call. The result is stated as a pipeline number, a revenue number, or a CAC payback figure, not impressions or MQLs. The timeframe matches your sales cycle.

Confirm at least 3 case studies pass all 3 filters before proceeding. If an agency cannot produce 3, they are a generalist with marketing copy, full stop.

Expected output: a shortlist of 3 or more relevant case studies per finalist with vertical, ACV band, cycle length, and sourced pipeline figure documented.

Step 2, Extract CRM-Backed Attribution Proof for Pipeline and CAC Payback

In the finalist round, require each agency to walk you through one client's attribution data live, with the CRM open. Salesforce or HubSpot CRM, plus Marketo or another marketing automation platform. The agency should show sourced pipeline by campaign, influenced pipeline by touch, and CAC payback calculated from actual closed-won deals. If they cannot show it in the CRM, it did not happen.

Attribution is where Tourists get caught. They will show you dashboards. You want the underlying report. A specific artifact looks like a Salesforce Campaigns report with Opportunity Amount, Close Date, Campaign Name, and Primary Campaign Source visible, or a HubSpot attribution export with the same fields. Redacted screenshots plus a live walkthrough are acceptable.

Ask four questions. What was CAC payback before you engaged, and what is it now? What percentage of closed-won revenue in the last 4 quarters touched a campaign you built? How did you define attribution windows for multi-quarter opportunities? Do you use multi-touch or single-touch attribution, and how do you handle dark-funnel touches across a 12-month cycle? Use your definitions, not theirs. See our marketing attribution glossary for the baseline.

When the agency says "we can't share CRM data," the acceptable answer is anonymized, aggregated, permissioned exports plus a reference confirmation. The non-negotiable fields are Opportunity Amount, Close Date, Campaign Name, and Primary Campaign Source. If they pitch AI as a shortcut around these fundamentals, disqualify them. This is the standard we apply to our own work at The Starr Conspiracy.

If you want a second set of eyes on a finalist's artifacts, we will pressure-test them with you.

Expected output: a verified attribution artifact for one comparable client, with field names and date ranges documented.

Step 3, Run Structured Reference Calls with Revenue-Accountable Buyers

Do not accept references from marketing managers. Request the VP of Marketing, CMO, or CRO who held the revenue number during the partnership. The person who signed invoices and defended spend to the board.

Use the same 5 questions on every call. What pipeline number did you commit to your board, and did the agency help you hit it? What did the agency get wrong, and how did they handle it? Would you hire them again at a 20% higher retainer? What is the one thing they do better than anyone else in your stack? Why did the partnership end, or why is it still going?

The answers you want are specific, numerical, and include at least one criticism. References who only offer praise are coached. Plan 30 minutes per call. Take notes verbatim.

Triangulate every reference claim against the CRM artifact from Step 2. If a reference cites a pipeline number that does not appear in the attribution export, ask why before moving on.

Expected output: 3 completed reference calls per finalist with a written summary scoring each on revenue accountability, candor, and consistency on a 1 to 5 scale.

Step 4, Read Client Retention Signals as Proxy Performance Data

Ask every finalist for average client tenure and 3-year retention rate. Cross-reference against logos on their website. If a logo appears on the site but the relationship ended 18 months ago, that is a tell.

Set internal vetting thresholds rather than chasing industry averages. Our internal operating benchmark at The Starr Conspiracy for specialist B2B agencies with genuine pipeline impact is 2.8 to 3.5 years of average client tenure, based on 25 years of operating in these verticals. Generalist shops run shorter in our experience across audits. Anything under 12 months across a portfolio signals execution problems. Anything over 5 years average can signal complacency.

Then the harder question. Which categories of clients have you walked away from in the last 24 months, and why? Agencies with real standards fire clients. The willingness to describe patterns of misalignment, without naming names, is one of the strongest positive signals in the vetting process. The Starr Conspiracy has walked away from engagements over revenue accountability, and we will describe the patterns without disclosing identities.

Confirm retention claims match logo history before proceeding.

Expected output: a retention scorecard per finalist scored 1 to 5.

Step 5, Stress-Test Channel-Specific Proof Across SEO, Content, ABM, and Email

Every agency claims integrated capabilities. Most have one strong channel and 4 they outsource or fake. Test each channel the partnership will actually use.

• Technical SEO: a redacted audit deliverable and resulting organic pipeline lift over 12 months. Confirm pipeline lift is shown in a CRM report, not a traffic chart.
• Content: the editorial calendar of a current client and attribution data tying pieces to opportunities. Confirm pieces map to Opportunity records by Primary Campaign Source.
• ABM: target account list methodology, the tech stack used, and meeting-to-opportunity conversion rate by tier. Confirm conversion rates are pulled from CRM, not the ABM platform's self-reported dashboard.
• Email: deliverability rates, list health practices, and revenue per send by segment. Confirm revenue is tied to closed-won, not opens.

Sequence channel proof to your sales cycle. Cycles over 12 months weight content and ABM heaviest. Cycles under 6 months weight email and paid acquisition. Deal sizes over $100K weight ABM and field marketing. Use our B2B ABM guide as a reference for what genuine ABM proof looks like versus account-list-plus-LinkedIn-ads theater.

Channel proof is where you confirm the agency is building a marketing system, not running disconnected campaigns. If they cannot show how the channels feed each other in a working account, they are stitching tactics, not building a system.

Expected output: a channel-by-channel score on a 1 to 5 scale with at least one verifiable artifact per channel.

How to Sequence These Procedures

The 5 procedures run in order, but where you start depends on your situation. Three decision rules.

If your sales cycle is under 6 months and your deal size is under $25K, start at Step 2 and weight CRM-backed attribution heaviest. Short cycles mean attribution windows are clean and excuses are thin. If your cycle is over 12 months and your deal size is over $100K, start at Step 1 and spend the most time on industry and cycle length relevance, because the wrong-vertical agency will burn 6 months learning your market on your dime. If you are replacing an incumbent that produced activity but not pipeline, start at Step 3 and pressure-test the specific failure mode that broke the last partnership.

Each step has a job. Step 1 gates whether the rest is worth running at all, so run it first and run it hard. Step 2 puts CRM proof under the case study story, which means the numbers either hold up or they don't. From there, Step 3 either corroborates or contradicts that proof. Whether the agency operates on standards or on revenue protection is exactly what Step 4 reveals. Step 5 confirms the channels add up to a system, and if they don't, you're done. Treat AI as augmentation of fundamentals, not a replacement for them. Do not let "AI-powered" claims substitute for proof.

Common Mistakes to Avoid

Accepting MQL volume as pipeline proof. In Step 2, the most common error is letting an agency substitute MQL counts for sourced pipeline or closed-won revenue. MQLs are a leading indicator at best, a vanity metric at worst. Force the conversation to opportunity stage and revenue. At The Starr Conspiracy, MQL counts are never an acceptable answer to a pipeline question.

Calling references from a list the agency curated without filtering. In Step 3, agencies offer their happiest 5 clients. Counter by requesting one reference from a partnership that ended, one from your specific vertical, and one at your deal size.

Treating retention rate as inherently positive. In Step 4, high retention can mask comfortable mediocrity. Pair retention data with the walk-away question.

Evaluating channel proof in isolation. In Step 5, buyers grade SEO, content, ABM, and email separately and miss the integration question. Ask how the channels feed each other in a working client account, with specifics.

Skipping the CFO validation step. Any CAC, payback, or ROI number an agency cites should be validated by your finance team against your own definitions. A short payback under one definition becomes a long payback under another.

The Bottom Line

Vetting a B2B marketing agency for pipeline and revenue impact is not a vibes exercise. It is a 5-procedure due-diligence workflow that takes 7 to 10 business days and produces a binary decision for your shortlist. This partner can prove revenue impact in my specific market, or they cannot. Run all 5 on every finalist. Do not negotiate the standard down because an agency is charming, recommended, or cheap. The cost of a wrong 12-month retainer is not the retainer, it is the quarter of missed pipeline that follows.

The Starr Conspiracy builds marketing systems that prove pipeline in the CRM. If you want that kind of partner, talk to us before procurement locks the SOW. We will walk you through the exact proof artifacts we expect you to demand of every finalist, including us, and the engagements we will refuse to take if the revenue accountability is not there.

Related Questions

How do I verify a B2B agency's CRM-backed attribution claims?

Require a live walkthrough of one client's CRM during the finalist round, under NDA. Ask for sourced pipeline by campaign, influenced pipeline by touch, and CAC payback calculated from closed-won deals. An agency that cannot produce these on screen in real time does not have the attribution discipline to drive revenue accountability in your account. See our marketing attribution glossary for definitions to align on before the call.

What should an agency provide under NDA when they cannot share raw CRM data?

Acceptable anonymized fields include Opportunity Amount banded into ranges, Close Date by quarter, Campaign Name, Primary Campaign Source, attribution model used, and CAC payback period. Non-negotiable: the agency must show field names and the originating system (Salesforce or HubSpot), and a named reference must confirm the figures. If they will not provide that minimum, disqualify.

What is a realistic CAC payback period for a B2B agency to deliver?

Use internal thresholds tied to your model rather than universal benchmarks. For SaaS with ACVs between $25K and $100K, treat under 12 months as strong. For enterprise fintech and manufacturing with longer cycles, set the threshold higher and validate against your finance team's definitions. Any agency claiming aggressive payback on enterprise deals is either citing a specific cohort or stretching definitions.

How long should I expect agency vetting to take?

A disciplined vetting cycle across 3 finalists runs 7 to 10 business days of focused work spread over 3 to 4 weeks. Shorter than that and you are skipping reference calls or channel-specific proof. Longer than that and the agencies you want will move on to other prospects.

Should I prioritize industry specialization or channel expertise?

For sales cycles over 12 months and committees of more than 5 people, industry specialization wins. The agency that knows your buying committee shortens the learning curve by months. For shorter cycles and simpler committees, channel expertise can outweigh vertical familiarity. Most B2B tech companies with complex sales motions are in the first category, which is why The Starr Conspiracy operates exclusively in those markets.