Low reply rates usually get blamed on copy. The team rewrites the opener, tests three subject lines, shortens the CTA, and the numbers barely move. Very often the problem sits upstream. Emails bounce, land with someone who left the company last spring, or reach the right person with nothing in the message that proves you know anything about them.
This guide is about that upstream part: the data you send from. It breaks response rates into three layers (deliverability, relevance and timing), shows what bad data does to each, and ends with a diagnostic loop for figuring out whether your next fix should be the list or the copy.
What "good" looks like
It helps to have a reference point. Instantly's Cold Email Benchmark Report 2026, based on campaigns sent through its platform during 2025, puts the average reply rate at 3.43%. The top quarter of senders reach 5.5% or more, and the top 10% clear 10.7%. The same report recommends keeping bounce rates under 2%.
If you're well under 3% and your bounce rate is above 2%, start with the data. If bounces are clean and replies are still low, the relevance and timing layers below are where to look next.
Layer 1: Deliverability
An email that bounces or goes to spam gets zero replies no matter what it says. Worse, it drags down every other email sent from that domain.
The data problems that hurt deliverability are boring and common:
Unverified addresses. Guessed patterns (first.last@) and old database records bounce. Mailbox providers read high bounce rates as a sign of a bad sender.
Contacts who changed jobs. Their old mailbox gets deleted, so an address that was valid six months ago hard-bounces today.
Catch-all domains. These servers accept mail for any address, so a verifier can't confirm the mailbox exists. Some will be real people, some will bounce later or go nowhere.
There's an infrastructure side too. Google's email sender guidelines require SPF or DKIM for everyone sending to Gmail and ask all senders to keep spam rates reported in Postmaster Tools below 0.3% (ideally under 0.1%). Domains sending more than 5,000 messages a day to Gmail need SPF, DKIM and DMARC. Set that up once, correctly, and move on.
The data fixes:
Verify every address right before a campaign starts, not when the list was built.
Remove invalid addresses. Send to catch-alls from a separate, smaller campaign so they can't wreck the main domain's numbers.
Re-enrich active lists on a schedule (monthly works for most teams) so job changes are caught before they turn into bounces.
For the sending-side problems that aren't about data, like warmup and inbox rotation, see our write-up on cold email deliverability issues.
Layer 2: Relevance
Once an email lands, the reader decides in a few seconds whether it's about them. That decision rests almost entirely on data you had (or didn't have) when you wrote it.
Relevance data comes in three kinds:
Role accuracy. The current title and team. Pitching sales tooling to someone who moved into partnerships reads as careless.
Company context. Headcount, industry, tech stack, business model. "Since you're on HubSpot" is only a good line if they're still on HubSpot.
Something specific and recent. A job post, a product launch, a new market, a podcast the prospect was on.
Compare two openers to a Head of Sales at a 90-person logistics software company:
"I noticed your company is growing and wanted to reach out."
"Saw you have four open AE roles in Chicago and just added a mid-market team page, so I'm guessing ramp time is on your mind."
The second one takes three data points: open roles by location, a website change and the prospect's current title. None of those require better writing. They require a table with the right columns filled in.
The trap is fake relevance. Tokens like {{company}} and {{industry}} dropped into a template look personalized to the sender and generic to the reader. If the detail would fit 500 other companies, it isn't relevance.
Layer 3: Timing
A perfectly relevant email can still arrive in a month when the prospect has no reason to act. Timing signals tell you when an account is more likely to be changing something.
Signal | Why it matters | Rough window to act |
|---|---|---|
Funding round | New budget and growth targets from the board | First 60 days or so |
New leader in your buyer's seat | New leaders often review tools and vendors early | First 90 days or so |
Hiring in the team you sell to | Growing teams hit process and tooling limits | While roles are open |
Tech stack change | They're already migrating, adjacent tools come up for review | Shortly after the change is detected |
Acquisition or expansion | Systems and teams get merged or rebuilt | Varies, watch for follow-on hiring |
The windows are rules of thumb, not laws. What matters more is speed: a signal is worth most in the first days after it appears, before every other vendor sends the same "congrats on the raise" email. We go deeper on building signal lists in our signal-based prospecting framework.
How the layers compound
Each layer multiplies the one before it, which is why fixing data often moves results more than any single copy test. Here's a worked example with made-up but realistic inputs for a 1,000-contact campaign:
Step | Unverified, stale list | Verified, refreshed, signal-filtered list |
|---|---|---|
Emails sent | 1,000 | 1,000 |
Delivered (after bounces) | 880 (12% bounce) | 985 (1.5% bounce) |
Reach the right, current person | 700 | 930 |
Reply rate among those | 1.5% (generic opener) | 4% (specific, timely opener) |
Replies | About 10 | About 37 |
Swap in your own numbers. The point is the shape: small losses at each step stack up, and the first list also damages the sending domain for next month's campaign.
Where to spend first
1. Verification before every send
The cheapest fix and the one with the fastest payoff. Verification runs at a fraction of a cent to a few cents per address depending on the tool and volume, and it protects every future campaign on that domain.
2. Scheduled re-enrichment of active lists
Rerun title, company and email enrichment on anyone in an active sequence or pipeline stage. Flag rows where the company domain changed: that's a job change, and sometimes a new opportunity at the new company.
3. Coverage from more than one provider
No single data provider has every contact. If you use one source, the people it misses never get emailed at all, or get a guessed address that bounces. A waterfall tries several providers in order and stops at the first valid result, so coverage goes up without paying every provider for every row. Check how many rows come back empty from your current source before and after.
4. Signals layered on top
Add columns for funding, hiring and leadership changes to your target account list, and route accounts to reps when a signal fires. This costs more setup than the first three, which is why it comes last.
Diagnose data before rewriting copy
The usual cycle when a campaign underperforms goes: rewrite the copy, resend, test subject lines, see a marginal change, decide cold email doesn't work. Try this order instead.
Check the bounce rate. Above 2%? Fix verification and list freshness before touching anything else.
Sample 30 contacts by hand. Open their LinkedIn profiles. If more than a few have a different title or company than your list says, re-enrich.
Read your "personalized" lines out of context. If a line could be sent to any company in the segment, you need better relevance data, not better phrasing.
Check timing. What share of the list had a signal in the last 60 to 90 days? If it's close to zero, you're emailing accounts with no reason to change.
Then work on copy. With clean data, copy tests actually tell you something, because you're no longer measuring bounces and wrong contacts. Our cold email copywriting rules are a good place to start.
FAQ
What is a good cold email response rate?
Per Instantly's 2026 benchmark report, the average is 3.43%, the top quarter of senders get 5.5% or more, and the top 10% exceed 10.7%. Your target depends on list size and how narrow the segment is. Small, tightly targeted campaigns should beat the average.
How often should I refresh outbound lists?
Verify right before each campaign. Re-enrich titles and companies monthly for contacts in active sequences, and before reusing any list that has sat for more than a couple of months.
Should I fix data or copy first?
Data, unless your bounce rate is already low and your contact sample checks out. Copy tests on a dirty list mostly measure noise.
Is it safe to send to catch-all addresses?
Partly. Some are real inboxes, some aren't, and the verifier can't tell which. Send to them from a separate campaign at lower volume and watch bounces closely.
Fix the list in one place
Databar puts the data side of this in one table: 160+ data providers, waterfalls for emails and phone numbers, verification, company and signal enrichment, and exports to HubSpot, Salesforce or your sending tool. You're charged for results, not empty lookups, and paid plans start at $99/month. Start a 14-day trial and run your current campaign list through it to see how many rows change.
Recent articles
See all
