I've been running RevOps for B2B sales teams for about 7 years. In that time I've personally signed off on enrichment decisions that wasted somewhere around $47,000 in budget. I keep a running doc of every one of those mistakes and I make new hires read it before they touch a vendor contract.
So when someone asks me what revenue operations teams should evaluate in a data enrichment company, my honest answer is: I don't know yet. Because the evaluation criteria change completely depending on what's actually broken in your pipeline.
Give me 30 seconds of context and I can give you a real answer. Without that context, anyone who hands you a checklist is selling something.
Figure out which bottleneck you're actually in
There are three scenarios I see over and over. Read these and be honest with yourself, because the last one is the hardest to admit.
- Scenario A — Your data rots. Records you bought 8 months ago bounce, job titles are stale, and SDRs keep emailing people who left two quarters ago. Your bottleneck is freshness.
- Scenario B — Your coverage is thin. You know exactly who you want to reach but you can't find them. Partial match rates, missing direct dials, gaps in specific regions or verticals. Your bottleneck is reach.
- Scenario C — You have records but no signal. Plenty of contacts, plenty of data, but your reps are guessing who to call first. Pipeline happens, it's just low quality. Your bottleneck is prioritization.
In my experience, most teams walk in saying they're Scenario A. When we dig in, roughly half are actually Scenario C. That's a big deal, because the fix is completely different.
Scenario A: Fix freshness before you fix volume
This is where I wasted my first $12,000. In 2020 I bought a 200,000-record list from a "premium" vendor. Felt great. Six months later, 45% of it was dead — bounced emails, wrong companies, the usual. I'd bought a snapshot, not a system.
What I should have evaluated for:
- Verification cadence. An email verification service is only as good as its last refresh. Ask: how often do you re-verify? What's your realistic deliverability rate over 90 days, not day one?
- Waterfall logic. Single-source enrichment will always have holes. Waterfall enrichment — querying multiple providers in sequence — closes more gaps and keeps records fresh longer.
- Decay tracking. Do they show you which fields decay fastest (titles and emails, usually) versus which stay stable (company domain, industry)?
The counterintuitive part: in Scenario A, buying more data almost always makes the problem worse. You end up with a bigger pile of rotting records. Fix enrichment cadence on your existing CRM first. You'll usually discover 30–40% of your problem disappears without a new contract.
Scenario B: Coverage is a stack problem, not a vendor problem
Scenario B usually shows up as a specific complaint: "we can't find enough VP-level contacts in mid-market fintech in Germany" or "we're getting 40% match rates on our target accounts." The temptation is to go find one vendor with a magic database.
That vendor doesn't exist. The strongest stacks I've seen use:
- One primary enrichment source for reliability
- One or two secondary sources that cover regional or vertical gaps
- A waterfall enrichment layer to query them in order and stop at the first hit
This is where outbound research tooling has genuinely changed. What used to be a manual analyst task — cross-referencing LinkedIn, ZoomInfo, Clearbit, whatever — is now something an AI agent can run at scale. But you need to evaluate how it decides. Ask: does the tool show you which source produced which field, so you can audit quality?
I have mixed feelings about agent-driven research, honestly. On one hand, it saves hours. On the other, I've seen agents confidently return garbage when the source was thin. The teams that get value from this keep a human-in-the-loop check on a sample every week. That's not a failure of the tool, that's just responsible operations.
Scenario C: You already have enough data — stop buying more
This is the painful one. If your reps have data but not direction, more enrichment is just more noise.
In 2023 we had about 90,000 records sitting unused because reps didn't know who to touch. I was about to buy another 50,000 enriched contacts when my director asked a simple question: "Why are we buying more of something we're not using?"
Dodged a bullet there. What we actually needed was intent data — signals that tell you who matters right now. Job changes, funding events, tech stack changes, content engagement, hiring surges. The right evaluation lens for Scenario C is:
- Signal quality and coverage. Which intent signals do you actually surface, and how fresh are they?
- Prioritization workflow. Can your SDRs see the ranked list without logging into a second tool? If it adds friction, they'll ignore it.
- Agent-native prospecting. Can an AI SDR agent take the signal, run the research, draft the outreach, and hand it to a rep for approval? That's the modern version of "workflow" — a research agent that produces a reviewable draft, not another dashboard to check.
The counterintuitive thing here: the best Scenario C solution often costs less than what you're already paying, because you cancel two of the tools it replaces. But that only works if you commit to one workflow. Teams that keep every tool "just in case" end up with the same prioritization problem, one subscription layer thicker.
How to know which scenario you're in
Here's the diagnostic I use with my team. Answer these three, in order.
- What's your 90-day deliverability on the email list you're currently running? If it's under ~85%, you're likely in Scenario A. Freshness is the problem.
- What's your match rate on a random sample of 100 target accounts? If it's under 60%, you're likely in Scenario B. Coverage is the problem.
- If I gave you 500 perfect records tomorrow from your ICP, would your reps know who to call first? If the answer is "they'd have to figure it out," you're in Scenario C.
One caveat: plenty of teams are in two scenarios at once. That's fine — just don't try to fix both in the same quarter. Pick the one bleeding the most pipeline and fix it first. The second fix is usually easier afterwards because the first one tells you what your data actually looks like.
And one more thing: this was all accurate as of Q1 2026. The GTM automation space is moving fast — new agentic options every few months — so verify current capabilities before you commit to a stack. Some of what I wrote three years ago in my mistake doc is already outdated, which is honestly the best sign that the industry is going somewhere.
The fundamentals haven't changed: know your bottleneck, match the tool to the bottleneck, and measure the outcome in pipeline, not in records purchased. That last part is where I lost the most money, and it's the easiest mistake to avoid if you ask the right questions before signing.
