How many people can you actually email?
Audience health separates your database from your audience. It counts who can be emailed at all, who among those is still listening, and who you are paying for and can never reach, using the fields your own tool already holds. Where you want to know how a given send performed rather than who is left to send to, read Which email sends worked, and is the list under strain? instead.
Before you start
Open Analysis, choose your tool, and click Audience health. It reads records already synced into your workspace, so a sync is what moves the figures. There is no Read the latest from button on this tab, and no read date above it.
No price appears anywhere
Your tool’s bill depends on your contract, and MarqIQ does not know it. This page says how many records sit behind that bill; you supply the rate.
Which fields it reads
Four readings, each found by what a field means in your workspace rather than by a fixed name. Up to four fields are taken for each of the first three, best filled first.
| Reading | What MarqIQ looks for |
|---|---|
| Opted out | A field whose words say unsubscribed from all, globally unsubscribed, do not email, email opt out, marketing suspended or address quarantined — and that is not a subscription-type, enrichment, SMS, tracking, reason or timestamp field |
| Bounced | A field whose words say hard bounce, invalid email, bad address or undeliverable, and that is not an enrichment, deprecated or soft-bounce field |
| Last engaged | A date field whose words tie an email open, click or reply to the words last, latest or recent |
| Billable | The field whose words say marketing contact status, marketable or marketing contact — the status itself, not the fields that describe it. Where several match, the one with the shortest label wins |
The method at the foot of the tab names every field it took for each reading, and prints “none found” where there was none. Read that before you read the counts.
A stored value is read as a person would read it: an empty string, false, 0, no, n, null, none and not_bounced all mean no. Anything else in an opt-out or bounce field means yes.
The denominator: every person, not a sample
This is the one analysis that reads your whole population however large it is. Rather than pulling records, it groups them in the database by the eight facts it judges on — an email address, an opt-out, a bounce, the billable flag, whether the person is a customer, whether engagement is recorded, whether it is over a year old, and whether the person has been in the tool long enough to judge. What comes back is one row per combination of those eight, each weighted by how many people share them, so the count is a few hundred rows however large the store. The figures therefore cover every person synced from that tool, inside whatever Created range you set, and none is sampled.
The four tiles at the top are People, Can be emailed, Gone quiet or never engaged and Paid for, unreachable. The headline gives the reachable count as a share, and adds the engagement figure when it can be judged.
What counts as reachable
An email address, no opt-out, and no bounce. Nothing else. Everyone else falls into one of three named groups, and the finding prints all three with their counts: people with no email address, people who opted out, and people whose address bounced.
That finding is raised once you have at least 100 people and a quarter of them cannot be emailed. It becomes a fault at 40%. The damage it states is the one leaders recognise: every plan that starts from “we have 82,000 contacts” starts from the wrong number, and the campaign target, the list-growth goal and the vendor tier are all judged against people who are not there.
Take care
Report audience size as the reachable count wherever the total is used today. Both numbers are on the tile row, so the substitution takes a minute and survives the first question it gets asked.
Who has stopped listening
Two groups, counted only among reachable people who are not customers.
- Gone quiet — the latest open, click or reply is more than 365 days old.
- Never engaged — no engagement recorded at all, and they have been in the tool for at least 180 days. A person who arrived last month is never blamed for silence.
The finding is raised when both groups together reach 30% of your reachable non-customers, on a base of at least 100 of them, and becomes a fault at 50%. It states its own denominator every time — “of 41,208 reachable people who are not customers” — because that number is much smaller than the database total and the two get confused.
Why customers are left out
A customer who ignores marketing email is still a customer, and is not a list problem. Excluding them is also what stops a business with a large installed base reading as though its list has collapsed.
The cost is stated in deliverability terms rather than in tidiness: mailbox providers decide whether your email reaches the inbox largely from how many recipients ignore it, so mailing the silent group lowers placement for the people who do respond. The bill is paid by your best audience.
People you pay for and can never reach
Where your tool holds a billable marketing-contact flag, MarqIQ counts the people that flag says are marketing contacts who have no address, have opted out, or have bounced. The finding appears from ten such people, and becomes a fault when they reach 3% of everyone synced.
It is the clearest waste on the page: the one thing a marketing contact can do is receive marketing email, and these cannot. The fix is to set them to non-marketing, and the finding says what that does not do — the records, their history and their deals all stay.
HubSpot is the clear case, because it bills by marketing contacts and carries the flag. On a tool with no such field this finding never appears, and its tile reads zero.
What you already have is checked before anything is advised
Before telling you to build a re-permission sequence, MarqIQ looks at the automations your workspace already runs whose names say they handle quiet people, and writes the advice around what it finds.
- Where one is switched on, it names it, gives how many people have been through it, and says that the count sitting here means it is not catching them — so the thing to check is what starts it and whether it ends by suppressing non-responders.
- Where one exists and is switched off, it says whether your team built it or the tool shipped it, and tells you to read it through before building anything new.
- Only where there is nothing at all does it describe a sequence to build: two or three emails asking whether they still want to hear from you, and suppression rather than deletion for anyone who does not answer.
The same is true of the billable finding, which checks whether an automation should already be setting opt-outs and bounces to non-marketing, and treats the count as evidence that it is not.
Checking a figure in your own tool
Every finding ends with a See it for yourself line naming the fields it read and telling you to build a list on the same conditions in your tool. The counts should be close. Where they are not, the usual cause is a different date window: the Created filter above the tabs scopes this analysis to people created in that period.
Where it stays blank, and why
| What is missing | What you see |
|---|---|
| No opt-out field found | The reachable count treats nobody as opted out, and the finding prints “read from no field, because none was found” in place of a field name |
| No bounce field found | The same, for bounces |
| No engagement date field | The headline says how many are still listening cannot be judged, and both quiet counts read zero |
| An engagement field that exists and holds nothing for anyone | The headline says so and adds that this usually means no marketing email has been sent from that tool. A field full of blanks means “not recorded”, never “not listening” |
| No billable flag | The Paid for, unreachable tile reads zero and that finding never appears |
| Under 100 people | The tiles and the headline, and neither of the two share-based findings |
What it will not do
- It changes nothing in your tool, and suppresses nobody.
- It never advises deleting a record. Suppression keeps the history and the deals.
- It quotes no price and no industry benchmark.
- It does not judge a customer on their silence.
If it did not work
The reachable count is far higher than you expected
Why: MarqIQ found no opt-out or bounce field, so nobody is being excluded on those grounds.
Fix: Read the method at the foot of the tab: it names the fields taken for each reading, or says none was found. Where your opt-out lives in a field whose label gives no clue, rename the label in your tool — the label is what is read.
Everyone shows as never engaged
Why: The engagement field exists and holds nothing, or the field found is not the one your tool actually stamps.
Fix: Check the field named in the method against a person you know opened an email this week. Where no marketing email has ever been sent from this tool, the empty field is correct and the counts should be read as “not recorded”.
The counts do not match a list built in your tool
Why: The Created range above the tabs scopes this analysis to people created in that period, and your list probably has no such limit.
Fix: Click Clear on the Created filter and compare again. The headline prints the population it used.
The figures have not moved since you cleaned the list
Why: This analysis reads records already synced into your workspace, and has no button of its own to re-read your tool.
Fix: Open Intelligence Studio, then Data sources, press Sync next to the tool, and reload the page when the sync finishes.
The Paid for, unreachable tile reads zero on a tool you know bills by contact
Why: No field in that tool was read as a billable marketing-contact status, or fewer than ten people match.
Fix: Check the method for the billable field it took. Where it says none was found and your tool does hold one, the label is not naming it — the reading looks for the words marketing contact status, marketable or marketing contact.
Next steps
Related articles
Was this article helpful?