What RevOps Teams Should Evaluate in a Contact List — After We Burned $84K on Bad Data

Our SDR team sent 4,800 LinkedIn connection requests last quarter. We got 41 replies. That's a 0.85% response rate — down from 14% two years ago on a smaller volume.

When I brought this to our VP of Sales, his first reaction was the same one every RevOps leader has: "The data's bad." And look — that's a reasonable first guess. Bad data is the easiest thing to blame. But after three months auditing our entire outbound stack, I can tell you it wasn't the data. It was how we were evaluating the data before we bought it.

If you're a RevOps team currently looking at contact list platforms — comparing okki-go account research against other options, evaluating LinkedIn prospecting tools, or just trying to build a framework that doesn't waste another budget cycle — I hope this saves you the $84,000 it cost us to learn the hard way.

What Everyone Thinks the Problem Is

When pipeline slows down, the diagnosis is always the same. We need more contacts. More emails. More LinkedIn touches. The problem gets framed as a volume issue, so the solution gets bought as a volume product.

Our RFP template asked the same three questions every time:

• How many contacts in your database?
• What's your cost per contact?
• How does your platform integrate with our stack?

These aren't bad questions. They're just incomplete — and they're incomplete in a way vendor sales teams have optimized around. One rep in 2023 pitched a "180-million-contact database." Those were his exact words. We signed two weeks later. Because what RevOps team says no to 180 million contacts?

Per FTC advertising guidelines, vendors are required to substantiate their claims with evidence. In practice, almost nobody in a procurement process actually asks for that evidence — myself included, until the invoices started piling up.

The Real Issue Nobody Puts on the RFP

Here's what took me 18 months and roughly 340 approved purchases to genuinely internalize: contact lists are not a volume product. They're a precision product. But almost nobody evaluates them that way, because precision is harder to put on a comparison spreadsheet than a headcount.

The 180-million-contact vendor didn't last six weeks with us. Our connect rate dropped from 18% to 6%. Same messaging, same ICP, same sequences. The only variable was the list. When I asked the vendor what percentage of those contacts had been active on LinkedIn in the last 90 days, I got a generic answer about "our proprietary freshness scoring." That was the first clue.

Three more months of digging surfaced three specific mistakes that I want to walk through, because they're the ones I see every other RevOps team making too.

Mistake 1: Buying coverage, not intent

When a vendor says "12 million contacts in your ICP," that number tells you nothing about whether any of them are in a buying window. It tells you nothing about hiring signals, tech stack changes, or LinkedIn engagement patterns. It's a ceiling, not a floor.

We spent an entire quarter prospecting into accounts where the contact was technically qualified — right title, right industry, right company size — but there was zero evidence they needed anything we sold. Reply rate: 0.6%.

What finally worked was layering intent signals on top of contact data. Job change alerts. Recent LinkedIn posts. Tech stack detections. This is what vendors mean when they talk about waterfall enrichment — basically, pulling contact and account signals from multiple sources instead of trusting a single database.

I don't have hard data on how much of our lift was intent versus better targeting versus pure luck. What I can say anecdotally is that once we started filtering lists by signals instead of just ICP fit, our meeting-booked rate nearly tripled within two months.

Mistake 2: Treating LinkedIn prospecting like email outreach

This one cost us one restricted LinkedIn account and a lot of SDR morale.

When we expanded into LinkedIn, we assumed we could take the same list, write shorter messages, and run the same sequence logic. That assumption was wrong in about four different ways. LinkedIn prospecting isn't about volume — it's about account research. The questions that matter before you send a LinkedIn message are different from the ones that matter before you send a cold email:

• What changed at this account in the last 30 days?
• Did they just hire into a function we serve?
• Are they posting about a problem we solve?
• Is there a competitor product showing up in job descriptions?

None of those come from a contact database. They come from research — the kind that used to take an SDR 20 minutes per account and now takes an AI agent about 30 seconds. The teams that figured this out before us were using tools that pulled together signals from LinkedIn, company websites, and job boards, and matched them against their ICP before a single message got sent.

There's a reason linkedin sales navigator automation exists as a category. Sales Navigator is a research tool — it shows you job changes, engagement, team growth. Automation on top of it is what makes research scalable. Automation without research is what gets accounts restricted. We learned that the hard way.

Mistake 3: Automating a broken process

The worst quarter we ever had was the one where we scaled our outreach by 4x. Same sequences, same lists, same messaging — just sent 4x as much. Reply rate didn't scale. It collapsed. We went from 14% down to 0.85% in nine months.

Why? Because we automated a process that hadn't been validated at small scale. When you send 200 messages to a mediocre list and get 12 replies, you think "the list is 6% good." When you send 4,800 messages to that same list, you don't get 288 replies — you get 41, because you've burned through every good-fit contact in the first 500 sends and then spent the rest of the quarter annoying the rest.

Volume doesn't fix bad targeting. It just makes it louder.

What It Actually Cost Us

I want to be specific about this, because "we wasted budget" is too easy to nod at without feeling it.

Across 2023 and 2024, here's what I can verify from our procurement records:

• $84,300 spent on prospecting data and enrichment subscriptions
• ~620 hours of SDR time spent working contacts that never should have been contacted
• 2 email domains with deliverability issues from over-sending to stale contacts
• 1 LinkedIn account restriction that took 11 days to resolve
• Two full quarters of garbage data fed into our CRM, which then required a clean-up project in Q1 2025

Add the CRM cleanup and the SDR ramp-back-up time, and the real number is closer to $140,000 in hard and soft costs. That's a full-time SDR salary. Gone.

So glad we didn't sign the two-year renewal on the original contract. Was literally one signature away from locking in another $96,000 on a vendor whose data hadn't worked for us in the first place.

The Evaluation Framework We Use Now

I wish I had tracked this stuff more carefully from the start. What I can share is the framework we've landed on after two years of trial and error. It has about five questions on it, and none of them are "how many contacts do you have?"

When RevOps teams ask me what they should evaluate in a contact list, this is the short version:

1. Signal density. What percentage of contacts have a fresh, verifiable signal in the last 90 days? (Not "last updated." Actual signal — job change, post, hire, tech stack shift.)
2. Enrichment sourcing. Is it one database, or a waterfall of multiple sources? Single-source databases are stale by default.
3. Account-level research capability. Can the platform tell you what's happening at the account, not just who the contact is?
4. Channel-native workflows. Does it support LinkedIn prospecting and email outreach as separate motions with separate logic? Or does it just reskin the same list twice?
5. Verification transparency. Can the vendor tell you exactly how they verify emails and phone numbers, and what their bounce rate is on your ICP specifically?

If you're looking at okki-go account research or the okki-go official website as part of your evaluation, make sure those five questions get asked. Doesn't matter which vendor you pick — the questions matter more than the brand on the invoice.

My experience is based on a mid-market B2B SaaS company with a 30-person sales org and a fairly defined ICP. If you're working with enterprise-length sales cycles or a far broader addressable market, your thresholds will look different.

An informed buyer asks better questions and makes faster decisions. That's the whole reason I wrote this up. The $84,000 was expensive. Passing the lesson along is free.