Settle on one source of truth for your metrics

The first step is a single table that shows what each tool reports for the same number over the same dates. Disagreements are impossible to settle when one person pulled last month in Pacific time and another pulled the last 30 days in UTC, so Kite fixes the date range and time zone before it reads anything.

Kite reads each connected tool the way an analyst would:

• Google Analytics. The sign-up key event and how many users triggered it, broken down by channel, source, and landing page.
• HubSpot. Contacts created in the period, with their original source, lifecycle stage, and how each record was created, such as a form, an import, or a salesperson.
• PostHog. The signup events your website or product sends, and how many distinct people sent them.
• Stripe. New customers and new subscriptions, including trials, along with whether each one went on to pay.
• Segment. How many events each source received each day and how many reached each destination, when Segment sits in the middle of your tracking.

If Kite built and hosts your website, its built-in analytics already records form submissions and conversions on the site, and Kite reads those too.

Here's what that table might look like for September. Google Analytics shows 512 sign-up events from 471 users. HubSpot shows 548 new contacts. Stripe shows 497 new customers. Those three numbers are the start of the investigation, not the answer.

Google Analytics and your CRM disagree because they measure different things in different ways. Google Analytics counts events that happen in a visitor's browser, and only when the browser lets it. Your CRM counts records, and a record can be created in ways that never touch your website. Kite measures each cause it can find, so the gap gets explained in numbers rather than guesses.

The most common causes fall into a few groups:

• Counting the same person more than once. A visitor who reloads the confirmation page can fire the sign-up event twice, and someone who signs up on a phone and again on a laptop can look like two users.
• People your browser tracking can't see. Cookie banners that visitors decline and ad blockers keep some signups out of Google Analytics entirely, while the CRM still gets the record.
• Records that never came from the website. Salespeople add contacts by hand, someone imports a list from an event, or a partner sends leads over, and all of them count as new contacts in the CRM.
• Test and internal accounts. Your own team signing up to check a new page shows up everywhere unless someone filters it out.
• Different time zones and dates. A signup at 11 p.m. on September 30 can land in September in one tool and October in another.
• Tracking that broke. A release that removed the tracking code from one page, or a renamed event that a tool now ignores, quietly drops signups from one tool from a certain date onward.

Kite checks each cause with the data that can confirm it. It compares sign-up events with users in Google Analytics to measure double counting. It reads how each HubSpot contact was created to separate form signups from imports and manual entries. It matches HubSpot contacts with Stripe customers on email address to find the people who appear in one but not the other. When Segment is connected, it reads daily counts per source to find the day a tracking problem started.

In the September example, the explanation might come out like this. Google Analytics counted 41 repeat events, which brings it to 471 people. HubSpot's 548 includes 37 contacts that sales added by hand and 9 from your own team, so 502 came from the signup form. Stripe's 497 is lower because 5 people who signed up never started a trial. Once each gap has a cause, the three numbers stop competing. They describe the same signups from three angles, and the remaining 31 people missing from Google Analytics are visitors whose browsers blocked tracking.

Related How-to pages

• How to find gaps in conversion tracking with Segment: Compare a count you trust, such as new accounts in your CRM, with what each Segment source receives and each destination records, and get every gap with its date, size, cause, and fix.

Each number should come from the tool closest to the thing it counts. Kite applies that rule to every metric your team reports and proposes one official tool for each, with the reason written beside it.

The usual result looks like this. Signups come from your product's own records, through PostHog or your CRM, because the account either exists or it doesn't. Leads come from your CRM, because that's where sales decides what counts as a lead. Paying customers and revenue come from Stripe, because a payment is the most reliable record of a sale. Traffic, channels, and landing pages come from Google Analytics, because it sees the visit before the signup.

That last point settles a lot of arguments. Google Analytics is rarely the right place to count signups, but it's usually the right place to ask which channel or page they came from. You can report 502 signups from the CRM and still use Google Analytics to say that 40 percent of tracked signups came from organic search.

Kite also writes the definition itself, in a sentence anyone on the team can check. For leads, that might be "A lead is a contact created through the website form or the demo booking page, with a work email, excluding existing customers and our own team." Getting your team to agree on how you count leads is mostly a matter of writing a sentence like that and having everyone read it before the next report goes out. Kite drafts it from what it found in your tools, and your team edits it until it's right.

An agreement that lives in someone's memory falls apart the first time that person is on vacation. Kite records each agreed metric in your team's knowledge base, the same shared place that holds your positioning, customers, and results from earlier work.

Each entry covers what someone would need to pull the number correctly:

• The name of the metric and its one-sentence definition.
• The official tool and the exact report, property, or filter to use.
• The time zone and how the dates are counted.
• What gets excluded, such as test accounts, internal users, or imported lists.
• How far the other tools usually sit from the official number, and why.

From then on, every report Kite writes uses the agreed number from the agreed tool and names its source. When someone asks Kite how many signups you had last quarter, the answer comes from the definition your team approved, not from whichever tool was easiest to open. If you correct Kite on how a number should be counted, it records that correction as a rule for future work too.

You can also ask Kite to publish the definitions as a page your team can open from a link, so new hires and agencies read the same rules you do.

The gap between your tools never disappears, but it should stay roughly the same size. Kite turns your first comparison into a baseline. If Google Analytics usually sees about 94 percent of the signups in your CRM and suddenly sees 70 percent, something changed, and it's almost always worth knowing what.

Kite can rerun the comparison on a schedule, such as the first Monday of each month, and report whether each gap held steady. When a gap moves, Kite looks for the cause the same way it did the first time. It checks for a date when one tool's count dropped, a new source of imported contacts, or a change in how a campaign was tagged. Then it tells you what changed and what to fix, before the next meeting turns into a debate about the data.

When you want the agreed numbers in front of everyone, Kite can build a dashboard at a stable link that shows each metric from its official tool, with its definition beside it.