How to audit B2B data Consultancy vendors for CRM data enrichment?

You’re three meetings into working out how to choose a data enrichment provider. Right now, you are comparing three decks and three accuracy percentages. However, you have no way to compare them, because none of the suppliers will say how that percentage was measured.

That’s the usual shape of a B2B data vendor evaluation: confident claims in a near-identical format, with nothing underneath to tell the suppliers apart. Some scorecard criteria are less useful at distinguishing credible supplier; only a handful of questions actually discriminate.

 

TL;DR

• Some vendor scorecard items are baseline requirements that credible suppliers should meet, so they don’t help you choose.
• Four questions to ask a data provider actually discriminate: sourcing, verification, ownership, and who does the work.
• Three apply whether you’re buying a platform or a service; the fourth applies to services only.
• A data enrichment pilot on your hardest records beats any questionnaire, if the scoring is fixed first.
• Ask what the pricing model actually rewards, and expect the same question asked of IMG (Intermedia Global Ltd).

 

How to Choose a Data Enrichment Provider: First, Platform or Service?

Before you get specific in your B2B data vendor evaluation, decide whether you want to go with a platform or a service.

A platform is data you license and query yourself. It is self-service, with subscription and usage-based pricing . A service is a supplier that builds the dataset to your spec and returns it enriched. Still deciding between the two? See data enrichment API vs data consultancy service.

Three of the four questions to ask a data provider below apply to both; the fourth, who does the work, applies to services only.

 

What Should You Ask A Data Enrichment Provider Before Signing?

Start by separating essential requirements from differentiators. Compliance, verification and CRM compatibility still need to be assessed, but you’re unlikely to see differentiation here e, so a B2B data vendor evaluation solely built on them produces near-identical scores, and the decision comes down to rapport.

The four below apply to platforms and services alike, except the fourth, which applies to services only. These questions are worth asking because 76% of surveyed CRM users and stakeholders said less than half of their organisation’s CRM data is accurate and complete (The State of CRM Data Management in 2025), and 81% of B2B buyers end up dissatisfied with the provider they picked (Forrester).

 

Question 1: Where Does The Data Actually Come From, and What Happens to Unmatched Records?

If a supplier won’t name sources, even broadly, rule them out on the spot, because you inherit their sourcing decisions, including compliance.

In 2025 alone, the ICO fined one company £90,000 for unlawful marketing calls where valid consent could not be demonstrated, and two energy firms £550,000 between them for unlawful automated marketing calls.IMG’s (Intermedia Global Ltd) enrichment approach draws on multiple sources that have been established as “trusted, compliant sources.” Clients should ask their provider what sourcing and compliance checks apply, and clarify the relevant lawful bases.

The sharper test is what happens to records nobody can match. A weak supplier quietly drops them and reports a strong match rate on whatever’s left. A strong one hands back an explicit not-found list, with reasons.

Buying a platform? Check whether its response distinguishes no match from low confidence, or just returns a blank. Buying a service? Ask to see a redacted exceptions output from a real engagement and not a demo.

 

Question 2: How Is Each Value Verified, and Is Confidence Recorded per Field?

Every supplier says it verifies its data, but the real test is how, and against what.

Is an email checked for deliverability, as only 83.5% of legitimate B2B sends reach the inbox, or just for format? Is a phone number validated by line type? Is a job title confirmed against a second source, or copied from whichever source returned it first?

Then ask whether confidence is recorded per field or per record. A single record-level score hides what you need to know: the same record can be solid on company data and shaky on the direct dial.

Ask when it was last checked, too, because a value checked eighteen months ago and one checked last week can both be filed as verified contact data.

 

Question 3: What Happens at the End of a Data Contract?

Three things need settling before you sign.

1. Who owns the enriched records

2. The rules

3. The suppression state, and is any of it only licensed?

Licensing is more common with platforms than buyers expect. For a service, is the method documented well enough that someone else could pick it up? And on exit, what format do you get everything back in, and how fast?

Ask IMG (Intermedia Global Ltd)to confirm the agreed delivery format and any usage or retention restrictions. We can supply the data “in the same format you supplied it” so uploading it into your CRM or system of record is simple. The goal is a capability you keep, not a dependency you renew.

 

Question 4: Who Does the Work, and Who Makes the Judgement Calls? (Services Only)

Now this question is pertinent if you are choosing services only. The people in the room during selection are frequently not the people doing the delivery.

So, ask directly: who’s assigned, what’s their experience with this kind of work, how much is subcontracted or offshored, and will you be told if that changes mid-engagement?

Then move on to who makes the judgment calls on ambiguous records? That’s rarely the person who pitched you. Get the named team written into the statement of work, not left in the proposal deck. If the answers here make you doubt the model altogether, it’s worth asking whether to hire an In-House Data Team vs Data Consultancy.

 

A Vendor Scorecard for B2B Data Vendor Evaluation

Procurement teams have used strong-answer/weak-answer scorecards for years. These ten are built specifically for how to choose a data enrichment provider, not generic supplier vetting.

 

Criterion	Strong answer	Weak answer	Applies to
Source transparency	Names sources and the lawful basis without being asked	“Multiple trusted sources,” no names, cites confidentiality	Both
Unmatched records	Explains no-match handling and provides an exceptions output where agreed	Silently drops them; quotes a match rate on what’s left	Both
Verification & confidence per field	States the method per field; confidence scored per field	“Verified” with no method; one score for the whole record	Both
Output ownership	Client has clearly documented rights to use, retain or export records, rules and suppression data, subject to licence and privacy obligations	Ownership is vague; some elements are only licensed	Both
Portability at exit	Named format and timeframe, confirmed in writing	“We’ll sort it when you leave”	Both
Delivery team & judgement calls	Named team in the SOW; a named owner for ambiguous records	“A dedicated team,” no names; unclear who decides	Service
Method documentation	Written up well enough for someone else to continue it	Lives in one person’s head	Service
Pilot terms	Comparable agreed test data, scoring fixed in advance; commercial terms clear	Will not agree a comparable test methodology	Both
Commercial incentive & declines	Names its model’s incentive and one thing it won’t do	Claims every capability; declines nothing	Both
Recency of verification	States when each value was last checked	“Verified” with no date attached	Both

 

How Do I Test A Data Supplier Before Committing?

Testing before you sign is something that procurement teams have always done. What’s missing from most advice is the design, and that’s where a data enrichment pilot either earns its keep or wastes a fortnight.

1. Include difficult records as well as a representative sample, so you can test both exception handling and overall performance. You’re buying help with the exceptions, not the average.
2. Fix the scoring before you see any results, including what a null result is worth, or you’ll rationalise whatever comes back.
3. Score exception handling as its own criterion, not folded into fill rate.
4. Give every supplier an identical file and identical instructions. A bake-off where each vendor scopes its own test tells you nothing.
5. Pay for it where a service is involved. Free pilots often get staffed by the pre-sales team, not delivery. A paid pilot may make it easier to assess the proposed delivery approach, but confirm who will actually perform the work.

 

Every Pricing Model Rewards Something, Including Ours

Every pricing model creates an incentive, and none is neutral.

• Per-record or per-credit pricing rewards volume. The supplier does better the bigger your list.
• Seat-based licensing rewards headcount, whether or not those seats get used.
• Retainers reward continuity, and can quietly reward doing less within it.
• Project fees reward finishing, and can penalise a supplier for surfacing extra work you genuinely need.

The incentive exists whether or not anyone acts on it, so price it into any B2B data vendor evaluation rather than assuming good faith. Ask the supplier to name its own model’s incentive and how it manages it.

Pricing models are changing. Seat-based pricing is already losing ground, down from 21% to 15% of companies in a year, with hybrid models rising from 27% to 41%. But the terms and usage assumptions of each proposal matter more than headline market averages. Ask what happens if usage, coverage needs or project scope change.

 

How Waterfall Enrichment Changes the Four Questions

Multi-source services like IMG (Intermedia Global Ltd) enrich records in sequence; one source fills what the last one missed, which changes how the four questions land.

Provenance becomes a per-field question, i.e., which source supplied this value, not just which supplier. A record remains unmatched after the sources agreed for that project have been checked without a usable result.

The exceptions output is whatever’s left after the last one. Ask how confidence and provenance are recorded for individual fields. This is how IMG enriches CRM data; ask any multi-source provider the same question: which source supplied which field, rather than assuming more sources automatically means a better result.

 

Red Flags: Where a Specific Answer Exists and a General One Is Offered

None of these disqualifies a supplier outright. They’re signals to probe further in any B2B data vendor evaluation, not automatic rejections. Listen for:

• A match rate quoted with no list behind it.
• Sources withheld as “commercial confidentiality.”
• Verification claimed with no method or date attached.
• No distinction drawn between no-match and low-confidence.
• No delivery team named, even when asked directly.
• Output handed over with no ownership statement.
• Every capability claimed; nothing ever declined.
• A pilot the supplier wants to scope itself.

If several of these show up together, go back to the questions to ask a data provider above. The scorecard already tells you what a strong answer sounds like.

 

Brief Against Your Campaigns, or the Scorecard Will Be Decided on Rapport

None of this works unless you brief the provider against the campaigns you’ve actually committed to: which roles you need to reach, which markets and segments, and what the output must support.

According to Forrester, B2B buying decisions involve around 13 people on average, and buying groups commonly run five to sixteen across several functions, as stated by Gartner. The right coverage requirement depends on the roles involved in your actual buying groups. A supplier who can’t help you reach those isn’t solving your real problem.

The outcomes to look for are clear:

• ABM programmes with detailed buying group data that teams can actually use
• Personalisation based on real data rather than AI working with incomplete records
• Audiences that work consistently across different platforms.

IMG (Intermedia Global Ltd) is a London-based B2B marketing consultancy, so this applies to us just as much as to any provider you may be considering. Our own marketing data enrichment services work follows the same four questions. We believe this shows that choosing a data enrichment provider and building a marketing data strategy are closely connected.

 

References:

1. Validity Inc. (2026, January 29). The State of CRM Data Management in 2025 – Validity. Validity. https://www.validity.com/resource-center/the-state-of-crm-data-management-in-2025/

2. Forrester. (2024, December 4). Forrester: The State of Business Buying, 2024. https://www.forrester.com/press-newsroom/forrester-the-state-of-business-buying-2024/

3. ICO. (n.d.). Compensation company fined £90,000 for unlawful marketing calls. https://ico.org.uk/about-the-ico/media-centre/news-and-blogs/2025/04/compensation-company-fined-90-000-for-unlawful-marketing-calls/

4. ICO. (n.d.-b). Warning over robo calls as energy firms fined half a million pounds for unlawful marketing calls. https://ico.org.uk/about-the-ico/media-centre/news-and-blogs/2025/09/warning-over-robo-calls-as-energy-firms-fined-half-a-million-pounds-for-unlawful-marketing-calls/

5. Validity. (2025). Email Deliverability 2025 Benchmark Report: An Analysis of Inbox and Spam Placement Rates Worldwide. https://www.validity.com/wp-content/uploads/2025/03/2025-Benchmark-Report-FINAL.pdf

6. Poyar, K. (2026, May 21). The state of B2B monetization in 2025. Kyle Poyar’s Growth Unhinged. https://www.growthunhinged.com/p/2025-state-of-b2b-monetization

7. Gartner Sales survey finds 74% of B2B buyer teams demonstrate “Unhealthy conflict” during the decision process. (2025, May 7). Gartner. Retrieved October 26, 2026, from https://www.gartner.com/en/newsroom/press-releases/2025-05-07-gartner-sales-survey-finds-74-percent-of-b2b-buyer-teams-demonstrate-unhealthy-conflict-during-the-decision-process