How to fix broken marketing attribution

Ad platforms see clicks. Web analytics sees visits and form fills. Neither knows which of those people became clients, or what they paid. That only lives in a CRM, the one place where a lead, a deal and its value sit on the same record.

Without a CRM, every channel can (theoretically) be measured up to the lead...but no further. You can't tell which brought in the most revenue.

Yes; they answer two different questions. Your CRM only knows about people who leave their details. Web analytics shows everyone else: which channels bring traffic, which pages people read, and where they drop off before converting.

GA4 and Adobe Analytics use cookies to recognize visitors. In the EU and UK they need consent to set them, so every visitor who declines the cookie banner is invisible to you (or those numbers get estimated rather than counted).

Cookieless analytics tools such as Matomo can count every visit, without relying on a consent banner.

Google, Meta and LinkedIn optimize toward whatever you tell them counts. If they hear about nothing, they optimize for clicks. If they hear about form fills from a pixel, they find more form fills, whether or not those "leads" ever buy.

Pixels also only see what happens in the browser, and are blind to everyone who blocks them or declines cookies. Conversions sent from your CRM are complete, and can speak to which leads turned into clients.

The further down the funnel, the better. A form fill is cheap to find and says little. A qualified lead says more, a closed client more still, and revenue tells the platform which buyers are worth the most.

However, signals like clients and revenue happen less often than form fills. Most teams send several steps, and let the platform learn from the deepest one it has enough of.

Click IDs (gclid, fbclid) are read by the platform that set them. Your CRM and most other tools can't decode them, so a lead arrives with a string of characters instead of a campaign.

UTMs are plain labels every tool can read. They're built from IDs in a tracking template (campaign ID, ad ID) rather than typed by hand, and they stay right even when a campaign is renamed. And (perhaps most importantly) nobody has to remember to add them.

Yep. Without them, a click from your newsletter or a LinkedIn post shows up as direct traffic or a referral (lies, damned lies), and the lead it brings is credited to nowhere in particular.

One rule covers most of it: every link you post or send that points to your own site gets utm_source, utm_medium and utm_campaign.

UTMs live in the address of the page someone lands on. As soon as they click to another page, a subdomain, or a separate booking or signup page, the address changes and the UTMs are gone (unless something on the site explicitly remembers them).

If the visitor converts two pages later, the form only sees a clean address, and the lead reaches your CRM without a source.

A form only sends the fields it has. If the UTMs aren't captured by hidden fields, or recorded by the form tool itself, the contact is created with a name and email...and no source.

Hidden fields and Zapier steps work sometimes, but you can't tell when they break. A new form without the fields, a renamed property, or a failed zap only get noticed when somebody realizes the reports aren't making any sense.

A booking tool is a form. If it isn't connected to your CRM (or only creates the meeting without the visit's source) then your best leads arrive with no channel attached.

Calendly, Cal.com and similar tools need their own integration to pass UTMs through, while a scheduler built into the CRM keeps the booking on the contact's record.

When leads are split between a CRM, another system and some spreadsheets, they can't be added up. CPL, cost per client, and revenue per channel all depend on counting every lead once (and only once), in the same place.

Spreadsheets are copied data, not connected data. They get exported from each platform, pasted, manually have their dates fixed, need to be rebuilt every week...and then the numbers are out of date by the time anyone reads them.

They also enable errors that are hard to catch. If someone leaves a filter on, or counts a month twice, the report still looks fine.

Most dashboard and connector tools pull numbers from each platform side by side: spend from the ad platforms, sessions from analytics, and deals from the CRM. Parallel, but not actually linked. If each deal isn't matched to the campaign that brought it, the dashboard can't say what a campaign earned (no matter how many platforms you connect).

Most teams stop at CPL, but your budget should be paying attention to the steps beyond that: cost per qualified lead, cost per client (CAC), and revenue per channel (or ROAS, return on ad spend).

Each step needs another link: spend, then leads, then the leads' status in the CRM, then deals and their value.

Each ad platform counts its own conversions, according to its own rules. They have different attribution windows and methodologies, and model (guess at) conversions for people they can't track. If you simply add up the numbers from all the platforms, you usually get more leads than you actually have.

Your CRM counts each real lead once. Spend should come from the ad platforms, but leads, clients and revenue should always come from your CRM.

Organic search has no campaign to tag, so it relies on the CRM recording the original source of each contact. Web analytics can show organic visits and even conversions, but not which of those people became clients.