How to Select a B2B Demand Generation Agency

How to Select a B2B Demand Generation Agency and Onboard Across Global Markets

To select a B2B demand generation agency and onboard one that rebuilds predictable pipeline across global markets, follow these five sequenced steps. You will need a defined revenue target, priority markets, pipeline data by source, executive sponsorship, and procurement access. This process takes 8 to 14 weeks. The Starr Conspiracy recommends gating each step on documented artifacts the board can review. For category context, see our demand generation glossary.

Step Summary Block

1. Build a geo-scoped agency shortlist tied to your priority markets.
2. Conduct a capability-fit assessment against your demand-state coverage gaps.
3. Scope ROI and unit economics before pricing conversations begin.
4. Structure the contract around outcomes, not retainer hours.
5. Execute a 90-day onboarding that ends with attributed pipeline.

Picture this: your CRO opens the board deck and the pipeline-to-plan gap is 38 percent. The directors want a name by Friday. Most agency searches die in that moment, because buyers collect logos before defining what the agency must actually do differently. That gap costs you a quarter, sometimes the CMO seat that came with it. Directories give you names. This gives you a board-defensible process, executed against five board-proof artifacts: Shortlist Doc, Scoring Matrix, ROI Scope, Outcome Contract Addendum, and 90-Day Onboarding Plan.

Prerequisites / What You Need Before Starting

Gather these inputs before running any step. Skipping them is the single biggest reason agency partnerships fail inside year one.

• A board-approved revenue or pipeline target for the next four quarters, broken out by region.
• Twelve months of pipeline data segmented by source, channel, and demand state.
• A named executive sponsor with authority to sign and to terminate at the 90-day mark.
• A current-state map of your martech stack: CRM, marketing automation platform (MAP), attribution, and enrichment tools.
• A written articulation of your buying committee, including dominant titles and named accounts in each priority market.
• A realistic budget range, not a ceiling, expressed as a percentage of the pipeline target.
• Procurement and legal availability inside the 8 to 14 week window. If legal cannot turn a master services agreement (MSA) in two weeks, the timeline slips.
• A security and data access plan: least-privilege provisioning, NDA, a signed data processing addendum (DPA), and an audit trail for every system the agency will touch. SOC 2 Type II or ISO 27001 attestation is the floor, not a tiebreaker.

If you cannot produce these artifacts in a week, pause the search and fix that first. Brand, message, and strategy are prerequisites for demand gen execution, especially when entering new geographies. See our GTM strategy framework before running Step 1. Use AI to accelerate analysis here, not to replace judgment.

Step 1. Build a Geo-Scoped Agency Shortlist

Start with geography, not capability. A full-service agency that prints pipeline in NYC will not automatically work in Tokyo or Stockholm. Build three separate shortlists if you are hiring across three regions, and accept that the answer may be two agencies, not one.

Decision fork: a single global agency is acceptable only when the buying committee profile, sales motion, and compliance regime are materially identical across your priority markets. If Japan and the Nordics diverge on any of those three, single-vendor is a trap. Push the call to Step 2 scoring.

For each priority market, pull candidates from review aggregators like B2B Marketing, capability directories such as the HubSpot Solutions Directory and Semrush Agency Partners, and your own peer network. Filter on four non-negotiables: documented B2B tech client work in that exact market, in-market or bilingual staff for non-English regions, compliance fluency (GDPR in the EU, PDPA in Singapore, APPI in Japan), and channel viability that often matches local buying dynamics (LinkedIn dominance in North America, events-and-relationships in Japan and DACH). Validate channel assumptions with two or three local sellers before finalizing the list.

Cap each regional shortlist at five names. Loop your regional sales leaders and country managers in here, not after; their veto comes anyway, and earlier is cheaper.

Verify: every shortlisted agency has named senior staff committed to the account, not just the pitch.

Expected outcome: a Shortlist Doc per region listing five agencies, four filter results, and named senior staff. The board can inspect this document and see exactly why each name made the list.

Gate to Step 2: the Shortlist Doc must be signed by the executive sponsor.

Step 2. Conduct a Capability-Fit Assessment

With the shortlist set, score each agency against the specific capabilities your pipeline gap requires. This is where most selection processes collapse into a beauty contest. Do not let it. Time-box this step to 10 business days.

Build a Scoring Matrix with three columns minimum: weighted criterion, evidence required, one-to-five score. Weight criteria against your demand-state coverage gaps. If your problem is late-stage conversion, weight ABM execution and sales enablement heavily. If your problem is category awareness in a new region, weight brand and content production. Require a written work sample, not a case study deck, for the top two weights. Adjust evidence requirements by region (localized creative for non-English markets, named regional accounts for ABM in DACH or Japan).

What we look for, pattern recognition from 25 years of running these searches:

• Operators name the practitioners on the account. Pitch shops name the partner and the AE.
• Operators show a real work sample inside seven days. Pitch shops ask for two more weeks.
• Operators discuss attribution before discussing creative. Pitch shops do the reverse.

If a finalist cannot produce a written work sample inside seven days, cut them. That seven-day rule is The Starr Conspiracy practitioner guidance, not an industry benchmark. Relax it only when the agency is mid-quarter close on a comparable engagement and offers a redacted version inside 10 days.

Verify: the Scoring Matrix is signed off by the executive sponsor and one peer outside marketing.

Artifact: a completed Scoring Matrix per region with finalists ranked, reducing forecast risk by binding the decision to evidence the board can audit.

Gate to Step 3: finalists confirmed in writing.

Step 3. Scope ROI and Unit Economics Before Pricing

Never let an agency quote a retainer before you have defined the unit economics the partnership must produce. Reverse the conversation. You tell them the cost-per-opportunity, pipeline-to-spend ratio, and payback period the board will accept. They tell you whether they can hit it.

Write a one-page ROI Scope with four fields: target pipeline contribution by quarter, maximum allowable blended cost per qualified opportunity (qualified opportunity meaning sales-accepted with a defined stage, value, and close date), the attribution model both sides will use, and the data-sharing cadence. Bring this document to finalist meetings. If an agency cannot map its proposed program to your unit economics in a working session, it is selling activity, not outcomes.

If you delay this step until after pricing, you negotiate against the agency's reporting template instead of your board's standard. Think of it as the difference between buying a car by horsepower and buying it by quarter-mile time. One is a spec sheet, the other is the outcome you actually need.

If your procurement or finance team is stalling on ROI scoping, that is a signal to bring in operators who do this weekly, not a reason to skip the step.

Verify: each finalist has returned a written response to the ROI Scope with assumptions and a sensitivity range. Anchor the model in your own pipeline attribution approach, not the agency's.

Sign-off: a signed ROI Scope document per finalist. This is the artifact that reduces forecast risk in front of the board.

Gate to Step 4: every finalist has returned a written response.

Step 4. Structure the Contract Around Outcomes

Most B2B agency contracts are written in hours and deliverables. That is how you end up paying for activity reports instead of pipeline. Restructure the agreement around three components: a base fee covering the operating system and team, a performance component tied to qualified pipeline or sourced revenue, and a 90-day termination option with a defined off-ramp.

The performance component should sit between 15 and 30 percent of total fees. This is The Starr Conspiracy practitioner guidance, not a sourced benchmark. Below 15 percent, the partner has no real skin in the game. Above 30 percent, you attract operators who game the metrics. If the agency's revenue mix is heavily enterprise with long cycles, push closer to 15. If the engagement is mid-market with quarterly close motions, push closer to 30. Define qualified-pipeline criteria jointly, in writing, before signing. Include a quarterly business review cadence with a named executive on both sides and a clear escalation path.

Objection handling: if procurement says outcome-based terms are not allowed, that is misaligned incentives dressed up as policy. Reframe to the CFO as forecast risk reduction and spend control, and procurement will follow. If the answer is still no, you have a procurement problem, not an agency problem.

Verify: legal and finance have reviewed outcome definitions and the data each side must produce to validate them.

Artifact: an Outcome Contract Addendum bound to the MSA, reducing spend risk by tying fees to pipeline the board can see.

Gate to Step 5: the addendum is countersigned.

Step 5. Execute a 90-Day Onboarding to Attributed Pipeline

The first 90 days decide everything. A good onboarding ends with the agency having sourced or influenced measurable pipeline. A bad one ends with a brand audit, a Miro board, and pageantry. We don't sell AI experiments here. We install operating systems that produce attributed pipeline. The agency you hire should do the same. By "operating system" we mean four named things: a reporting cadence with owners, attribution rules with definitions, campaign ops with SLAs, and governance with an escalation path.

Run the 90 days in four phases:

• Weeks 1 to 3: Provision least-privilege access to CRM, MAP, ad platforms, and analytics. Confirm DPA and audit trail are live. Install UTM governance. Define lead lifecycle stages and the SDR-to-marketing SLA. Document a baseline for every metric the partnership will be judged on.
• Weeks 4 to 6: Launch a live program in one priority market, even if small, to prove the operating motion. Stand up the weekly operating cadence between agency leads and your team. Localization QA runs in-market, regional approvals are owned by named country managers, and time-zone cadence is set (who attends, when, what decisions get made).
• Weeks 7 to 10: Expand to remaining markets. Deliver the first attribution readout against the Step 3 model.
• Weeks 11 to 13: Run the 90-day review against contracted outcomes, with a written go or no-go decision from the executive sponsor.

What good looks like at day 90 (in our searches, not a universal benchmark): meetings booked up two to three times baseline in the pilot market, sourced opportunities tracking against the ROI Scope's quarterly target, pipeline influenced visible in CRM with attribution rules both sides agree on. AI augments the analysis cadence; it does not replace the practitioners running the program.

Verify: the agency has produced attributed pipeline against the Step 3 model, the data both sides report matches inside an agreed variance, and the security audit trail shows no anomalies.

Sign-off: a 90-Day Onboarding Plan with a signed go or no-go from the executive sponsor. The Starr Conspiracy treats this artifact as the only meaningful proof point. Everything before it is busywork.

Gate to year-two planning: the go or no-go is in writing.

Common Mistakes to Avoid

Treating the shortlist as the answer. In Step 1, buyers pull a ranked list from a directory and skip geographic and compliance filters. The result is an agency that has never operated in your priority market quoting confidently about it. Filter on in-market proof first.

Scoring on chemistry instead of capability. In Step 2, finalist meetings turn into rapport sessions and the Scoring Matrix gets quietly abandoned. Force the written work sample inside seven days. Chemistry is the tiebreaker, not the criterion.

Letting the agency define ROI. In Step 3, buyers accept the agency's standard reporting framework instead of imposing their own unit economics. You get metrics that flatter the agency, not metrics that satisfy the board.

Signing a retainer-hours contract. In Step 4, procurement defaults to time-and-materials because it is familiar. Outcome-based structures are harder to negotiate and orders of magnitude more useful when the program underperforms.

Skipping the 90-day go or no-go. In Step 5, the executive sponsor lets the review slip because the agency is likable and the program is busy. Six months later you are renegotiating instead of exiting, and the board is asking why pipeline has not moved. The Starr Conspiracy has watched this kill more partnerships than every other mistake combined.

The Bottom Line

Selecting a B2B demand generation agency is a procurement procedure, not a vibe check. The buyers who rebuild predictable pipeline run the five steps above as a gated sequence with five board-proof artifacts: Shortlist Doc, Scoring Matrix, ROI Scope, Outcome Contract Addendum, and 90-Day Onboarding Plan.

If you only do three things: enforce the seven-day work sample in Step 2, write the ROI Scope before any pricing conversation in Step 3, and hold the 90-day go or no-go in Step 5. Those three gates filter out non-executing vendors without you having to name them.

If your board is asking where the pipeline is and next quarter's meeting is inside six weeks, start Step 1 this week. Talk to The Starr Conspiracy. We will run a 30-minute pipeline baseline review, then build the Shortlist Doc and Scoring Matrix with your team inside a two-week sprint. We help multi-market revenue leaders produce the five board-proof artifacts and install the operating cadence that delivers attributed pipeline. We don't sell AI experiments. We build marketing systems that actually work.

Related Questions

How long should the full agency selection process take?

Eight to 14 weeks from prerequisites through signed contract, plus a 90-day onboarding before you judge results. Compressing below six weeks almost always means skipping the capability assessment or the ROI scoping, and those are the two stages that predict year-one success. See our GTM strategy framework for broader sequencing.

Should we hire one global agency or multiple regional partners?

Hire multiple regional partners by default, and consolidate only when you can prove the buying committee and sales motion are identical across markets. If priority markets share a buying committee profile and a sales motion, one partner with regional staff is more efficient. If Japan and the Nordics require fundamentally different GTM approaches, two specialist partners coordinated through your team will outperform one generalist trying to span both.

What is the right performance fee percentage in an outcome-based contract?

Between 15 and 30 percent of total fees as The Starr Conspiracy practitioner guidance, with the qualified-pipeline definition agreed in writing before signing. Below 15 percent the agency has no meaningful exposure to the outcome. Above 30 percent you risk attracting operators who optimize for the metric rather than the business.

What should we include in an agency RFP for multiple regions?

Issue one RFP shell with region-specific addenda, not one global RFP. The shell defines unit economics, attribution model, qualified-opportunity definition, and contract structure. The regional addendum defines compliance requirements, language and localization expectations, channel mix, named accounts, and in-market reference requirements. Score each region independently against your demand state coverage gaps.

How do we know an agency can actually execute in our market and not just claim it?

Demand three references with comparable ACV, sales cycle length, and regional footprint, then ask each reference what the agency does badly. Vague answers mean the reference is coached. Specific, unflattering answers mean the reference is real, which is the signal you want.