An open rate that counts security scanner hits is a number your client can't use. Say you report 60% opens on a campaign that produced two sales. Sooner or later the client asks why, and you have nothing to tell them. This guide shows you how to track real human email open rates with a process you can repeat for every client. You'll record the inflated figures, learn what machine opens look like, clean the list, send through reporting that leaves scanners out, and judge each campaign on human opens read next to clicks and bookings.
The timing is good this week. It's late January 2024, the holiday sends are finished and January is one of the busiest months of the year for email, so your baseline resets anyway. From February, the large mailbox providers also start enforcing stricter rules for bulk senders. Those rules cover authentication, easy unsubscribes and spam complaint levels. If you're going to look hard at your numbers, do it now, before February's changes and your corrections arrive in the same report.
Here's the honest part up front. Your corrected open rate will come out lower than the old one, and it's an estimate, not a perfect count. It's still the first open rate you can make decisions with.
By the end you will report opens a client can act on, and you need only four things to start
When you finish, each client report carries three things. You get a corrected open rate based on human opens divided by delivered emails, with clicks and bookings shown beside it so every number backs up the others. You also get audience segments built on what real people did, so your next send goes to readers rather than machines.
You need four things before you start.
- Access to past campaign reports, with raw opens, clicks and send times you can export.
- A client list you can export, including bounces and unsubscribes.
- A sending platform that filters scanner opens out of its reports. We'll use SendTo throughout.
- A client who has already had reports from you, so there's a before to compare with the after.
Budget an afternoon for your first client. Most of that goes on the baseline and the scanner review in Steps 1 and 2. After that, each campaign takes well under an hour, because the cleaning and filtering run on their own.
Step 1: Write down the inflated numbers before you change anything
Export the last three to five campaigns for the client. Pull raw opens, unique opens, clicks, delivered count and the time each campaign went out. Put them in a spreadsheet with one row per campaign. Don't tidy anything up yet, because this record of what you used to report is the whole point.
You'll see open rates that look too good next to weak clicks. A B2B list showing 55% to 65% opens with a 1% to 2% click rate is the usual pattern. Here's a table you can copy. The after column stays empty until Step 4.
| Measure | Before (raw, last 4 campaigns) | After (first filtered campaign) |
|---|---|---|
| Sent | 10,000 | 10,000 |
| Delivered | 9,850 | 9,400 |
| Open rate | 61% | 25% |
| Click rate (of delivered) | 1.9% | 1.8% |
| Click to open ratio | 3% | 7% |
| Bookings or replies | 14 | 15 |
Look at the bottom three rows. Clicks and bookings barely move. The open rate drops by more than half, and the click to open ratio more than doubles. That pattern tells you the old opens were mostly machines, and it gives you a figure you can point to when you explain the change.
Step 2: Learn the fingerprints that give a scanner away
Pick one recent campaign and open its raw activity log, the event-by-event list rather than the summary. You're checking for three patterns. The first is opens logged within a few seconds of delivery. The second is every link in the email clicked at the same moment by the same recipient, including the unsubscribe link and the footer address. The third is bursts of activity from one corporate domain, with dozens of addresses at the same company opening inside the same minute.
Corporate security filters cause most of this. Before a message reaches the inbox, the filter fetches the images and follows every link to check for malware. Each image fetch counts as an open, and each link it follows counts as a click. We cover the mechanics in detail in our piece on how security scanners distort open tracking.
Some mail apps cause a second kind of false open. They preload every image in a message as soon as it arrives, as a privacy feature, whether or not anyone reads it. The tracking pixel loads with them, so the recipient shows as opened even if they delete the email unread. These opens are harder to spot, because they don't come in corporate bursts. They show up as an even spread of opens soon after delivery with no clicks behind them.
You'll see whole B2B domains opening at once. One company with forty addresses on the list can produce forty opens in eight seconds. That isn't a team reading together. It's one filter checking forty copies.
Step 3: Clean the list so bounces and dead addresses stop skewing the denominator
Scanners inflate the top of the open rate. Dead addresses distort the bottom. Fix both, or the corrected number is still wrong. Verify addresses before every send and apply suppression when you import, so hard bounces, past unsubscribes and known bad addresses never reach a campaign.
In SendTo, emails are verified before sending. List imports map spreadsheet columns automatically, pick up interrupted uploads where they stopped, and apply platform-wide suppression. A client's messy export becomes a sendable list without you fixing columns by hand. If you're unsure whether this is worth doing on B2B lists, we've written up the case for automated email list cleaning.
Your delivered count will come out smaller, and that's what you want. The open rate formula is simple. Divide unique human opens by delivered emails, then multiply by 100. Use delivered, not sent, because an email that bounced had no chance of being opened, and counting it treats a dead address as someone who chose not to read.
Here's how that works with the table above. Of 10,000 sent, 9,400 were delivered and 2,350 humans opened. That's 2,350 divided by 9,400, which is 25%. Divide by sent instead and you get 23.5%, and the gap grows as the list gets dirtier.
Step 4: Send through reporting that removes scanner opens before anyone sees them
Run the client's next campaign through SendTo. Its reports leave out security scanner opens and show human opens, clicks, bookings and variants. The same figures appear in the client portal and in your partner view, so you and the client are reading one set of numbers. You don't need to set up a filter or add tags yourself. The report you open already has the scanner opens removed.
Two other parts of the platform keep this test fair. New sending domains get automated warmup, and campaign volume is spread across several brand-matched sending domains, so no single domain carries the full load. That matters here. If a fall in opens came from a cold domain landing in spam, you'd blame the filtering for a deliverability problem. Our piece on how automated domain warmup prevents new accounts sending to spam explains why that first fortnight matters.
You'll see a lower open rate and a click rate close to what it was before. Fill in the after column of your Step 1 table. If opens fell from around 60% to the mid-20s while clicks held at roughly 2% of delivered, the filtering did its job. Your real audience hasn't shrunk. You're just no longer counting machines.
The common mistake is comparing this campaign with the old raw figures as if they measured the same thing. They don't. The old number counted people plus scanners plus preloaded images. The new one counts people. From here on, compare filtered campaigns only with other filtered campaigns, and keep the old figures for one job, explaining the change.
Step 5: Track real human email open rates alongside clicks and bookings, never on their own
Build every client report around three numbers, in this order. Human opens, then clicks, then bookings or replies. If the client sells by phone or by form, count whatever starts a real conversation. Putting them top to bottom turns the report into a funnel, and the client reads how people moved through it, not one headline figure.
You should see each stage make sense next to the one above it. Take the after column from Step 1. There were 9,400 delivered, 2,350 human opens, about 170 clicks and 15 bookings. Each stage loses people at a rate you can explain. The old report showed about 6,000 opens leading to 14 bookings, and nobody can plan around a stage that loses 99% of its people.
So should you stop reporting opens altogether? No. Keep them, but move them down the page. Clicks and bookings lead, and opens explain them. A low open rate with a healthy click to open ratio points at the subject line or the send time. A healthy open rate with weak clicks points at the content or the offer. You can only tell those two apart when you track real human email open rates and read them beside clicks.
Opens still settle A/B subject line tests, even with some noise left in
This is the part the formula guides leave out. Split a list at random, send both variants at the same time, and scanners hit variant A and variant B at roughly the same rate. Filters don't read subject lines. They check every message that arrives. The noise lands evenly on both sides, so the gap between the variants comes mostly from people. SendTo's reports show variants side by side with human opens for each, which makes the comparison quicker still.
There are two conditions. The split has to be random and the send times have to match, or one variant lands in a different inbox mix. On small lists a gap of two or three points is just noise, so wait until clicks agree before you call a winner.
Step 6: Build segments on clicks and unclicked opens, not on raw opens
Open the client's audience in SendTo and build three segments from the last few filtered campaigns. One holds the people who clicked, one holds the people who opened but didn't click, and one holds the non-openers. SendTo builds segments from exactly those behaviours in past campaigns. We walk through the setup in our guide to segments based on campaign interaction history.
Your engaged segment will come out smaller than the old openers list. On the client in our table, four raw campaigns gave an openers list of around 5,800. Clickers plus human unclicked openers came to about 2,600. That smaller group contains nearly everyone who books. Send the offer to the clickers. Give the unclicked openers a shorter email or a different angle, because they saw the message and the content didn't move them. Try the non-openers on a new subject line and send time, and retire them if they still don't respond after a few sends.
The mistake here is re-sending to openers who were really scanners. Each of those sends goes to a security filter. It costs volume, does nothing for the client and props up the next report with fake engagement. It matters more from February, when the bulk sender rules make poor engagement and complaints harder to ignore.

Step 7: Show clients the corrected numbers before they find the gap themselves
Open the client portal view and put your Step 1 before and after table beside the new report. Add two sentences of explanation. Send it a day before the report goes out, or open your next call with it. Don't let the client find the drop on their own.
From this month your open rate leaves out the security filters that open every email automatically, so it will read around 25% instead of 60%. Your clicks and bookings haven't changed, and this new number is the one we'll use to decide what to send next.
You should see the client accept the lower number, because you raised it first and showed them that clicks and bookings held steady. Will clients think performance dropped? Only if they spot it before you mention it. The usual mistake is switching to the new numbers quietly. A 60% that turns into 25% with no explanation looks like you lost half their audience in a month.
This conversation also helps you keep the client. An agency that corrects its own flattering figures is one a client believes next time, and trust is what keeps a monthly retainer paid. Since the portal runs under your brand, the honest report comes from you too. If you're turning advisory work into a platform line, we cover how agencies turn strategy consulting into recurring software revenue.
How to check it worked, and what to do when the numbers still look wrong
After two or three filtered campaigns, run three checks. Each one takes a couple of minutes in the report and the activity log.
- The click to open ratio has risen. In our example it went from 3% to 7%, because the denominator now counts people.
- Opens no longer arrive in a burst in the first minute after delivery. They spread out over hours, the way people actually read email.
- Segment sizes make sense. Your clickers and unclicked openers together come to a fraction of the list, not two thirds of it.
Three problems and where to look for each
If opens are still high with almost no clicks, sort the activity by recipient domain. A single domain often accounts for the gap, with a company filter that behaves unusually. Judge that domain on clicks and replies only, and ask whether it belongs on the list at all.
If opens fell sharply and clicks fell with them, filtering isn't the cause. It removes machine activity, not human clicks. Look at deliverability instead. Check bounce levels, whether a new domain is still warming up, and whether authentication is in order before the February rules arrive. Filtering pushes opens down and leaves clicks alone. A deliverability problem pushes both down.
If a client's B2B list is mostly one company, the open rate will never be clean, because one filter shapes the whole send. Report clicks and replies, and leave opens in as a footnote. This matters far more for B2B lists than for consumer ones. Business domains sit behind link-checking filters, while consumer lists mostly show the image-preloading kind of false open, which spreads evenly and distorts less.
Affiliates who can explain this problem earn from every agency they bring in
If you write about software, consult for agencies or run a channel for marketers, this subject is easy to recommend from. Agency owners search for answers when a client asks why 60% opens produced two sales. A walkthrough like this one answers that question. A platform that reports human opens, clicks and bookings in a client portal under the agency's own brand is the obvious next step to point them to.
SendTo's affiliate programme pays you 20% of the partner subscription each agency you refer pays SendTo, every month that agency pays it. You never get a share of what the agency's own clients pay the agency. Attribution lasts 90 days from the first click, and the first touch wins. Commission is paid monthly by PayPal. When your readers ask what the platform costs, send them to the pricing page for current plans rather than quoting a figure that may change.
Write about honest open rates, send agencies to the platform that reports them, and earn 20% of every referred agency's partner subscription for as long as it pays.
Join the SendTo affiliate programme


