We Wasted $18K on Prospecting Tools. Here's the Checklist I Wish I'd Had.

2026-08-24 · Julian Hartwell

If you've ever had to present a sales tech budget to a CFO and watch them slowly page through your line items, you know the mix of hope and dread that comes with tool evaluation. I've been on that ride three times in six years as a RevOps lead. Twice, I got it wrong enough that I'm still feeling the effects. This is the checklist I wish someone had handed me before the first purchase order.

When This Checklist Helps

Use this if you're evaluating sales intelligence platforms, intent data providers, or power dialers for a B2B sales team. It's built for RevOps professionals and SDR leaders who need to make a defensible decision in the next 30–45 days — not a theoretical one. It took me three failed evaluations and roughly $18K in wasted budget to get here. I'm not saying this list is perfect. But it would have caught every single mistake I made.

Quick context on where I'm coming from: I've been handling sales stack evaluation and revenue operations for a B2B SaaS company for about six years. In my first year (2022), I made the classic mistake of choosing a platform based on database size instead of fit with our actual buyers. Then in September 2023, I renewed an intent data subscription that our SDRs had already quietly stopped opening. After the third failed evaluation in Q1 2024, I sat down and wrote the process I use now. It's saved us from at least two bad purchases since. Trust me on this one — the steps below are the ones I would put in front of any team about to spend money on prospecting tools.

Step 1: Lock Down Your ICP Parameters Before You Open a Single Demo

I started my first evaluation convinced I knew our ideal customer profile. I was about 60% wrong.

The question I wrestled with was simple to ask and surprisingly hard to answer: what should revenue operations teams evaluate in ideal customer profile? I now use four buckets:

  • Firmographic signals: company size (headcount and revenue band), industry sub-segment, funding stage, region. Don't just list "software" — specify exactly which kind of software companies you win in.
  • Technographic signals: which tools your best customers already run, whether they're on a competitor's product, and whether they have the engineering capacity your implementation depends on.
  • Behavioral signals: hiring patterns (are they growing the exact team you sell to?), tech stack changes, and budget signals like job posts for the role that would own your category.
  • Pain-point signals: the problems your champions mention in reviews, conference talks, and support threads.

My mistake in 2022: I wrote down three descriptors — company size, industry, region — and called it an ICP. It made a clean slide. But it produced prospect lists that looked nothing like our best customers. We spent a full quarter hammering accounts that had the right firmographic fit but zero urgency. Nobody wanted what we sold, and we only found that out after three months of demos.

Here's the practical check: pull the criteria you're considering into a search inside any sales intelligence platform. Does the result look like your 20 best customers? If not, keep refining. Your ICP needs at least two dimensions that are observable in a prospecting database — otherwise you'll never be able to scale it beyond a slide deck.

Step 2: Test Intent Data Against Your Closed-Won Deals

This is the step that saved me the most money, and it's the one most teams skip.

When you look at an intent data platform, the demo will show you dashboards, topic clusters, and lookalike audiences. It all looks deeply technical. Don't buy from a demo. Run a backward test instead.

Take your last 20 closed-won deals. Upload those accounts into any intent data platform you're evaluating — they'll all give you a trial. Then ask three questions:

  • How many of those 20 accounts showed intent signals in the 30–90 days before they became a lead?
  • Did the signals align with the actual buying topic, or was it a vague "someone at this company searched for something adjacent to what you sell" blip?
  • Does the platform give you account-level intent, contact-level intent, or both?

My September 2023 disaster: I evaluated three intent data providers back-to-back. I compared their dashboards instead of comparing them against revenue. The one I picked looked great in the trial — and I renewed it for six months before realizing our SDRs had stopped opening the alerts by week three. Why? The signals were too generic. Our team got 40+ alerts a day and learned to ignore all of them.

It's tempting to think intent data is intent data. But the difference between "this company is researching your category" and "this company is actively comparing vendors exactly like you" is the difference between a deal that takes four months to close and one that takes four weeks. What most people don't realize is that intent data quality varies far more by your specific segment than by vendor brand. The only way to know is to test it on your own won deals.

Step 3: Stress-Test the Power Dialer With Your Actual Workflow

A power dialer feels like a low-stakes decision. It's just a dialer, right? Not exactly. If your SDR team lives in it for 40 hours a week, small frictions become massive productivity losses.

Here's what I check now:

  • CRM call logging: can reps log a call in one click, or do they have to type fields manually?
  • Compliance features: do dropped calls route to voicemail properly, or do they end up as missed calls on real people's mobile phones?
  • Local presence: does it actually work, or do your prospects see an unfamiliar area code and ignore it?
  • Pacing: is there an annoying delay between calls that wastes 30+ seconds per hour per rep?
  • Click-to-dial from your sales intelligence data: can your SDRs see a verified number and dial it in one action without switching tabs?

The thing nobody tells you: multi-line dialers (the "power" part) can burn your team if the data feeding them is bad. If your sales intelligence tool gives you phone numbers with poor accuracy, your SDRs are dialing disconnected lines and blaming the dialer. It's not the dialer. It's the data underneath.

I went back and forth on the Lusha vs UpCell question for our SDR team for about two weeks. UpCell looked like the more cost-conscious option on paper, and a couple of our reps liked the interface. But when we ran the numbers, the cheaper license got eaten by wasted dials — too many dead numbers and a few disconnected mobiles that should have been flagged as invalid. Lusha wasn't dramatically more expensive, but the per-contact accuracy and email verification built into the records meant far fewer dials went nowhere. Bottom line: don't pick a dialer without testing the data that feeds it, because the dialer is only as smart as the numbers it's given.

Step 4: Check Data Depth on Your Actual Personas — Not the Demo's Poster Child

This is where I see the simplification fallacy play out constantly. "We compared five platforms and Lusha vs ZoomInfo looked similar in coverage." Fine. On which accounts? If your evaluation list is made of Salesforce and Microsoft, every vendor will look identical. That's not the point.

Here's the test I run now:

  1. Take 10 accounts from your top 20 ICP fits.
  2. Pick 3 contact titles per account — say, VP Sales, VP RevOps, and Director of Sales Development.
  3. Check each platform: how many of those 30 contacts have a complete, verified email and a direct phone number?
  4. Actually dial the phone numbers. Does the direct line connect to the person, or does it route to a switchboard?
  5. Send a test email to the verified addresses. Does it bounce? If the platform says "verified," does that mean the mailbox was confirmed, or was it just syntax-checked?

In my 2023 evaluation, ZoomInfo had bigger overall volume at almost every data point. Nobody disputes that. But on our actual target accounts, the coverage gap between Lusha and ZoomInfo was far smaller than the price difference — and Lusha's records were complete in the exact titles we care about. ZoomInfo's huge database was impressive; it was also full of companies we would never call. Never expected that to be the deciding factor, but it was.

I should say the honest flip side too: a colleague at a recruiting firm swears by ZoomInfo because their market is so broad that they need that massive warehouse. The point isn't which platform is "better." The point is that volume claims are worthless if coverage on YOUR ICP is thin. Put another way: a database with 200 million contacts is no more useful to you than a database with 50 million if your buyers all live in the smaller one.

Step 5: Check Integration and Workflow Fit Before You Sign Anything

This sounds obvious, but it's where I lost the most time — and it's a harder failure to spot because it doesn't show up until after you've committed.

I once signed a six-figure contract with a platform that scored highest in our accuracy tests but required SDRs to open a separate tab, export lists, and upload CSVs to get data into the CRM. The team hated it, so they stopped using it. We paid for a stack that got maybe 3% utilization before I finally killed it at renewal.

What I evaluate now:

  • Does it integrate natively with your CRM, or does it need middleware?
  • Can your reps enrich a lead in one click from inside the CRM?
  • Is there an API your RevOps team can use for automated workflows? Lusha's agent-native API is built around this — you can connect it to automation tools and trigger enrichment in real time.
  • Does the data flow back into your CRM when a rep updates a record?
  • If you're building AI-assisted prospecting, can the platform expose data in a way those agents can actually consume?

We moved almost all of our prospecting workflows into tools that connect without a development sprint. The "best data" that requires manual handling is worse than slightly-noisier data that flows automatically.

Step 6: Model the Total Cost — Including the Cost of Bad Data

The last mistake I'll admit to: I chose a tool with a slightly lower per-user price but forgot that the data credits were significantly tighter. Two months in, we exhausted our monthly email verification quota, and our SDRs couldn't enrich or dial until the next cycle. Thousands of dollars in salaries sat idle because I saved $60 per user per month.

When you compare platforms, model the full picture:

  • License cost × number of users
  • Data credits — how many per month, and what happens when you hit the cap? Does enrichment freeze, or do you get charged overages?
  • Email verification — unlimited or per-record cost?
  • API call costs if you're automating enrichment
  • Setup and migration costs
  • Expected renewal price. If I remember correctly, one of my renewals came in about 15% higher than year one — though I might be misremembering the exact figure. Either way, get the year-two price in writing before you sign.

Here's the piece that ties back to brand, and it's why I'm willing to spend on quality: every email that bounces, every number that's dead, every list that needs manual cleanup — that's not just a productivity cost. It's your company sending broken messages to buyers who might otherwise have taken you seriously. The $50 per month you save on a questionable data source gets spent twice on corrections, and the third cost is the impression your SDRs leave on prospects who received a poorly targeted or undeliverable outreach. So glad I stopped treating data quality as just a line item, because the cheap option was quietly damaging the one asset we needed most: credibility.

Common Mistakes I Keep Seeing

If you take nothing else from this, avoid these five:

  1. Starting the tool search before the ICP is done. I did this, and let the vendor's pitch shape our ICP. We ended up targeting accounts the tool was good at, not accounts we were good at winning.
  2. Comparing data volumes instead of coverage on your accounts. Volume is a neck-snapping demo metric. Coverage on your segment is the operational reality.
  3. Skipping the intent data backward test. If a platform can't flag signals from your last 20 customers, it won't spot your next 20.
  4. Not checking verification dates. Here's something vendors won't tell you: "verified" can mean verified eleven months ago. If the data hasn't been re-checked since your prospect changed roles, you're working with dead records.
  5. Valuing the dialer over the data feeding it. A power dialer is only as good as the numbers it dials.

There's no such thing as a no-brainer choice in sales intelligence. There's only the choice you make after testing the data, the workflow, and the total cost against your actual revenue motion. Take it from someone who learned the hard way: the 30 days you spend running a proper evaluation costs a fraction of what the wrong six-figure contract costs. Every time.