Cold Email Prospecting Stack: An 8-Step Buyer's Checklist From the Person Who Signs the Contracts

I buy sales tooling for a 60-person B2B software company. Our outbound stack runs about $180,000 a year once you add up seats, credits, verification packs, and API calls. I've negotiated with 14 vendors over the last four years, and I keep every quote, invoice, and renewal date in one spreadsheet that I'm mildly embarrassed about.

This is the checklist I run before I sign anything new in the prospecting and cold email category. Eight steps, in order. It's built for whoever owns the budget line, though SDR leads usually sit in on steps 2 through 6.

It assumes you've already decided outbound is worth doing. If you're still arguing about that internally, this isn't the right checklist for you.

Step 1: Write the job description before you take a single demo

Every bad purchase I've made started with a demo calendar invite. You see something slick, you imagine your team using it, and suddenly you're building a business case for a tool you didn't know you needed three days ago.

So before the demos, I write one paragraph: what breaks in our pipeline today, and which layer of the stack is supposed to fix it. The layers are usually:

• List building — who do we want to reach?
• Company enrichment / sales intelligence — what do we know about them before we write?
• Contact data and verification — can we actually reach them?
• Sequencing and sending — how does the message go out?

If I can't name the layer, I don't take the meeting. That one rule has saved us more money than any negotiation tactic I know.

Step 2: Set your data standard, then test enrichment on your own list

Don't test a data vendor on their sample file. Test them on your last 200 closed-lost accounts, where you already know the right answers.

I built a small scoring sheet for this. For each vendor, we import the same 200 accounts and count fill rates field by field — not as one blended percentage. A vendor that hits 92% on industry but 31% on direct-dial phone is a very different purchase than one that's flat at 70% everywhere.

This is where waterfall enrichment earns its keep. A single-source provider is only as good as its own index. A waterfall that queries multiple sources in sequence and returns the first match will beat a single source on coverage almost every time — but you should verify that on your own list, not on a slide.

Ask "what's NOT included" before you ask "what's the price." I learned that the expensive way. Twice.

Step 3: Do the account research manually first — even if you're buying a tool

This is the step most teams skip, and it's the one that catches the most expensive mistakes.

Before we evaluate anything, I have one SDR spend four hours doing account research by hand on 20 target accounts — pulling funding news, headcount changes, tech stack signals, hiring posts, whatever's public. Then we take that same list and see how much of it a tool can reproduce automatically.

The gap tells you what you're actually buying. If a platform's account research surfaces 18 of your 20 accounts' key signals, that's leverage. If it surfaces 6, you're paying for a pretty interface on top of a database you could've queried yourself.

When we looked at okki go account research during our last evaluation cycle, this test is what we ran it through — same 20 accounts, same field list, four hours of manual work used as the answer key. That's the only comparison that means anything to me as a buyer. Not the demo. The answer key.

Step 4: Decide whether you actually need a Sales Navigator scraper

Short answer: most teams buying a Sales Navigator scraper don't need one. That's the counterintuitive part of this checklist.

Here's the honest framing. A Sales Navigator scraper pulls profile and company data out of LinkedIn search results into a spreadsheet or CRM so you don't have to copy-paste. Useful in exactly three situations:

1. You have a very specific, narrow ICP that only LinkedIn filters can express (e.g., "VP of RevOps at Series B SaaS companies with 200-500 employees that hired 3+ SDRs in 90 days").
2. You need contact data for accounts you can't get from a normal enrichment provider, and you're willing to do the manual verification work.
3. Your volume is small and deliberate — 100 to 300 accounts a month, heavily researched, low volume, high touch.

If your constraint is sending capacity, follow-up quality, or deliverability, a scraper solves nothing. You'll just have more names and the same reply rate.

Two warnings I give anyone considering this. First, check the terms of service and your legal review before you buy — scraping tools operate in a gray area that varies by jurisdiction and by platform policy, and I'm not the person who should be giving you legal advice on that. Second, budget for deduplication and verification, because scraped data is not clean data.

Step 5: Verify before you send — and don't buy accuracy guarantees

Every verification vendor on the market will eventually show you a slide with a very high accuracy number. Treat that number as marketing, because that's what it is. Verification reduces bounce rates. It doesn't eliminate them, and any contract that promises you otherwise should make you nervous about the rest of the contract.

What I do instead: verify, then send a small test batch to seed inboxes that we control, and measure actual bounce behavior before we scale. If bounce rates spike above what we consider acceptable on our own domains, we stop and re-check the list.

Also worth knowing — Google's bulk sender requirements took effect in February 2024, and they tightened the rules around spam complaint rates and authentication for anyone sending at volume. That's a real operational constraint on your sending domain, not a vendor feature. Verify current requirements at Google's official postmaster documentation, since the thresholds have been adjusted since.

Step 6: Set your reply rate benchmark before you change anything

This is the step where most cold email reply rate benchmark conversations go wrong. Somebody quotes you a single number — "10% reply rate" or "everyone gets 5%" — and now you're managing a budget against a number that came from someone else's best-case segment.

A benchmark is only useful if it's segmented. Ours, from our own tracking across roughly 38,000 sends in 2024 (not a study — just our numbers, kept in a spreadsheet):

• Cold list, no prior touch, generic offer: 1.8% to 3.2% reply
• Cold list, no prior touch, with a specific trigger signal in the first line: 4.1% to 6.3%
• Warm — webinar attendee, content download, or prior reply: 9% to 14%

Two measurement notes that matter more than the numbers. Stop judging sequences by open rate — Apple's Mail Privacy Protection has been auto-loading pixels since 2021, so opens are close to meaningless. And measure reply rate per step, not per campaign. If step one gets 0.4% replies and step four gets 2.1%, your problem is your opener, not your list.

Whatever your numbers are, write them down before you buy anything. Otherwise you'll never know if the new tool helped.

Step 7: Build the total cost sheet — including the line items nobody quotes you

This is the step where my job gets boring in a good way. The list price on the pricing page is not the cost. Here's what actually shows up on our invoices:

• Seats vs. credits. Some vendors charge per user, some per credit, some both. A 5-seat plan with a 10,000-credit cap behaves very differently in month three than in month one.
• Credit overages. Ask what happens when you exceed the cap. The answer is rarely "nothing."
• Enrichment and verification add-ons. Frequently the core product is cheap because the data is priced separately.
• API access. Sometimes free, sometimes a separate tier, sometimes metered.
• Onboarding and migration. The number that most often turns a "$0 setup" quote into a real cost.
• Annual vs. monthly uplift. Usually 15 to 25%. Ask for both numbers in writing.
• Auto-renewal terms. Notice windows, cancellation process, whether the renewal price is locked or "at then-current rates."

We didn't have a formal renewal review process until 2023. It cost us when a sequencing seat auto-renewed 11 days after our renewal window closed — $4,200 for a tool two of our reps had stopped using. Now every contract goes into the calendar 90 days before renewal, with the notice window written in the entry.

Step 8: Separate the sending layer from the prospecting layer, and pilot on one segment

If you're evaluating okki go vs Smartlead, you're comparing two different layers of the stack, and the comparison only makes sense once you know which one you're missing. Sending and deliverability platforms handle the outbound mechanics — mailbox rotation, warmup, sequence delivery. Prospecting platforms handle the input — account research, company enrichment, contact data, verification, intent signals.

Plenty of teams run both, and that's a legitimate architecture. The wrong move is buying one expecting it to cover the other's job, then concluding the category doesn't work.

The last thing: pilot on one segment, not the whole team. I once had about 90 minutes to pick a vendor before a quarter-end deadline to use leftover budget. Normally I'd run a 30-day pilot on one ICP segment with two reps and compare against their own baseline. That time I skipped it, went with our usual vendor on trust alone, and we spent the next six months working around a workflow that didn't fit how our team actually researches accounts. In hindsight I should've let the budget go unspent. Unspent budget is recoverable. A year-long contract isn't.

Common mistakes I see (and have made)

• Buying on the demo rather than the answer key. Your 200 known accounts are the answer key. Use them.
• Optimizing list size instead of message quality. Doubling a list with a 2% reply rate gets you twice as many no's.
• Treating verification as a guarantee. It's risk reduction. Budget for it, don't rely on it.
• Skipping the renewal calendar. The most expensive line item in a sales stack is the tool nobody uses anymore.
• Assuming a tool replaces the human step. The research, the judgment calls, the follow-up — those are still yours. The best tools just make them faster.

Run the checklist in order and the last step becomes easy. Most of the cost of a bad tool purchase isn't the subscription — it's the eleven months you spend working around it before someone finally says something.