Which form fills are real, and which are junk?
Form fills answers three questions about the number every marketing report rests on: how many fills you really had, how many of them were junk and why, and what a real fill turned into in your tool. This article explains the tab and what each figure is counted over. For the exact rules behind a junk label, and how to check one yourself, read What counts as a junk form fill? instead.
Before you start
Open Analysis, choose your tool, and click Form fills. On HubSpot and Marketo the tab always runs, and until the script is live it says so: “No form fills have been captured yet. They appear here once the MarqIQ script is live on a page with a form.” On every other tool the tab runs only once the script has captured at least one fill; before that its button is drawn in grey with a dashed border, and opening it gives one sentence headed Not something this tool can do. The sentence is quoted in full under The forms your tool holds below.
Where the fills come from
MarqIQ’s own captures, recorded by the script on your website. They are not read back from your tool, which is why a workspace with a full CRM and no script sees an empty tab. On HubSpot and Marketo the tool’s own forms appear in a separate section above, and that section is read from the tool.
The four numbers at the top
Read these before anything else, because each later rate is measured on one of them.
| Tile | What it holds |
|---|---|
| Captured | Every fill MarqIQ recorded in the period. Where the read hit its cap the tile reads “the most recent; older fills are not shown” instead |
| Recorded twice | The same person filling the same form again within two minutes. Counted once, and set aside before any judgement |
| Junk | The count and the share, each with its reason listed below |
| Real fills | What every rate on this page is measured on |
The analysis reads up to 60,000 captured fills, most recent first. That is the one cap on this tab, and it is the exception to how every other analysis reads a large store — everywhere else a cap turns into an even spread across the whole history.
Take care
Where repeats reach a tenth of what was captured, What to do next says so, because any report counting raw fills is overstating volume by about that much. The figures on this page already have them removed; the figures in your tool probably do not.
What the junk is, and where it comes from
Seven reasons, each with its count, a plain explanation of what it means, and up to three masked examples such as jo**@mailinator.com — enough to recognise, not enough to copy. The full rules are in What counts as a junk form fill?*
Below that, the card headed Where the junk comes from ranks your sources by the share of their fills that are junk.
- A source is listed only with at least 30 fills in the period. Twelve sources at most.
- A source is called out only when its junk share is at least one and a half times your overall junk share, or at most half of it. The card prints your overall share in its own note so the comparison is visible.
- Fills with no source recorded are kept as their own row, named No source recorded, rather than being dropped.
The action that follows names the money: where the worst source is paid traffic, you are paying for the junk, and the fix is the targeting or a verification step on the forms it lands on.
Your forms, and what each one produces
One row per form, busiest first, up to 40 of them. A form is shown by the name your tool gives it; where the tool has no name for it, by the page it sits on — never by an identifier.
| Column | What it is measured on |
|---|---|
| Form | The name your tool gives the form, or the page it sits on |
| Fills | Every fill captured on that form, before repeats and junk are removed |
| Junk | That form’s junk share, of its own fills |
| Found in HubSpot | Of its real fills carrying an email, the share matched to a person synced from your tool. The column is named after whichever tool you are on |
| Went on to buy | Of the same real fills with an email, the share who are a customer or sit on a sales deal |
| Mostly from | The commonest source among its real fills |
| Last fill | The date of the most recent one |
A form carries a gone quiet tag beside its name when it took at least ten fills and none in the last 60 days. That usually means the page stopped loading the form, or the script came off it.
Fills that never reach your tool
The follow-through rates compare what MarqIQ captured against what your tool holds, matched on the email address. Three thresholds decide how much the tab will say.
- At least ten real fills with an email address before the two rates are shown at all.
- At least 30 matched before the tab will diagnose it. Below that you get the rates and no diagnosis.
- At least ten fills per form before that form is named in the diagnosis.
Where under 70% of real fills can be found in your tool, the action distinguishes two shapes and says which you have. Where some forms are above 85% and others below 50%, it names the failing ones: those forms are not connected to your tool, or their email field is not mapped. Where every form with enough fills is low, it says the cause is shared — the sync from your forms to your tool, not one form’s settings.
On a very large store
Matching normally tries the address exactly as typed and again in lower case. Above 200,000 records in that tool the second pass is skipped, and a line above the tab says the match was on the exact address only. A tool storing “Ana@Acme.io” against a form that captured “ana@acme.io” then reads as unmatched.
Personal mailboxes
A personal mailbox such as Gmail is not junk and is never counted as such. Where your business profile says anything other than business-to-consumer, the share of real fills using one is reported on its own, as a fact. It is raised as an action from 40%, because for a business selling to companies it makes people hard to match to an account.
The card needs at least 30 real fills with an email before it appears. A business-to-consumer profile never sees it, because there the personal mailbox is the point.
Month by month
Twelve months at most, each split into real fills and junk. A month where junk jumps usually lines up with a campaign or a form change, and reading the two together is what tells them apart.
The forms your tool holds
On HubSpot and Marketo a second section sits above MarqIQ’s captures, built from the tool’s own forms: how many people each form brought in, how many became customers, and how many live forms nobody has filled. It has its own tiles — Forms that brought people in, People brought in, Became customers and Live forms nobody filled — its own findings, and its own Read the latest from button. Its reading of your tool is kept for a day, and the date is printed above it.
On every other tool that section is absent, and what you see next depends on whether the script has captured anything. With no captures the whole tab is muted and carries one sentence in its place: MarqIQ cannot read that tool’s own forms yet, so it cannot say which form brought someone in — and it names the two things that cover the same ground, Source truth, which reads the fields your records already carry about where each person came from, and the fills the tracking script captures, which appear here as soon as the script is live. Once the script has captured a fill the tab opens normally, and the section about your tool’s own forms is absent with nothing said about it.
Take care
A Marketo instance with 52 approved forms and no fill activity is told exactly that. “52 live forms brought nobody in” is a finding about your forms, not a failure of the analysis, and the two are never dressed up as each other.
What the figures are measured on
- Every rate on the tab is over real fills, after repeats and junk are removed. The four tiles give you all three populations so any figure can be converted back.
- The follow-through rates are over real fills that carry an email address, which is smaller again. The line under the forms table prints that denominator in full — “72% of real fills can be found in HubSpot (1,340 of 1,861)”. In the table itself each form’s two counts are on the cell’s hover, not on the page, so quote the line and not a column.
- The Created range on this tab is the date a fill was submitted, not the date a person was created. The line above the cards reads “fills submitted” and then your period.
- A source or a form is only compared when it has at least 30 fills, and only when the difference is both larger than chance could explain and large enough to act on.
What it will not do
- It changes nothing in your tool and deletes no fill.
- It never calls a fill junk for an unusual real name, or for a personal mailbox.
- It does not judge a fill it has no signals for. The automated reason is read only for fills captured since fill-quality signals were switched on for your workspace, and that setting is off until an admin turns it on.
- It reads no fills from your tool’s database. What is here is what the script captured.
If it did not work
The tab says no form fills have been captured yet
Why: The tracking script is not live on a page with a form, or it is live and the form is not being seen.
Fix: Install the tracking script, then check a form yourself and reload this tab. The script is installed but nothing is arriving covers the usual causes, including a snippet in a block that never runs.
The fill count is far lower than your tool reports
Why: Your tool counts submissions from every source, including forms MarqIQ’s script never sees — an embedded form on a page without the script, an offline import, or a partner’s landing page.
Fix: Compare form by form rather than in total. A form missing from this list entirely is a page the script is not on.
A form appears named after a page instead of by its name
Why: Your tool has no name for it — usually a plain HTML form, or one built outside the tool.
Fix: Nothing has to change for the counting to work. Track plain HTML forms covers giving those forms an identity so they carry a name here.
The junk share jumped this month
Why: A campaign started pointing at a form, or a form changed and lost whatever was keeping automated fills out.
Fix: Read Month by month against Where the junk comes from. A single source at several times your usual junk share is the first place to look.
Only a small share of fills can be found in your tool, and no form is named
Why: Fewer than 30 real fills with an email have been matched, so the tab shows the rates and holds back the diagnosis.
Fix: Widen the Created range and look again. Below 30 the card says the loss is spread thinly and to compare once each form has at least ten fills with an email.
A fill you know is real is counted as junk
Why: One of the seven reasons matched it — most often a flood it arrived inside, or a name read as filler.
Fix: Open What counts as a junk form fill?, which gives the exact rule for each reason and how to check a sample before acting on the number.
Next steps
Related articles
Was this article helpful?