Conversion: what separates macro from micro, and what changed with GA4

A conversion isn't a fixed action that means the same thing on every site. It's whatever action you decide counts as success inside your analytics tool: a purchase on an online store, a quote request on a services site, a newsletter sign-up on a media site, or a case study download on a B2B site. None of that is a "conversion" by default. It becomes one the day someone sets it up and marks it that way.

The metric that summarizes this activity is the conversion rate: the number of conversions divided by the number of sessions or users, expressed as a percentage. If a store gets a thousand sessions in a week and logs twenty purchases, that week's conversion rate is two percent. That figure means little on its own: a fashion ecommerce site and a B2B consultancy start from completely different typical conversion rates, because the cost and commitment of the final action aren't comparable.

In Google Analytics 4, conversion tracking runs on an event-based model: every interaction on a site or app, a page view, a click, a form submission, a purchase, gets recorded as an event with its own name and parameters. Any event can later become what GA4 currently calls a key event, the term that replaced "conversion" inside the admin interface, though "conversion" is still the everyday word in the industry and also lives inside Google Ads, where the two concepts connect.

Before configuring anything, it's worth separating a conversion from nearby metrics it gets confused with. It isn't the same as an impression or a session, and it isn't a synonym for traffic: page visits can double without moving the conversion rate a single point if the problem sits in the offer itself or in process friction, not in visitor volume. And unlike the customer journey, which describes the full path a person takes across channels, a conversion is only the endpoint of one concrete, measurable action.

An ecommerce site with a short purchase cycle can afford to look almost exclusively at the macro conversion: the purchase happens in the same session or a few days later, well inside the window GA4 tracks reliably. The problem shows up in businesses with a long sales cycle, where the macro conversion takes weeks or months and often happens outside the analytics platform itself.

A typical B2B sales cycle doesn't fit inside one session, or inside a short conversion funnel. A visitor lands on a blog post, comes back weeks later to download a guide, signs up for a webinar, requests a demo, has several conversations with sales, and months later signs a contract that often gets closed over the phone or logged straight into the CRM, with that last step never touching the website or firing a single measurable event in GA4.

If reporting only counts the macro conversion, almost everything that happens before it disappears from view. A channel producing excellent content and attracting the right people can show "zero conversions" for months, not because it isn't working, but because the actual result of that work, the signed contract, falls outside the attribution window and often outside the analytics platform altogether. With that incomplete read, it's easy to cut budget from the exact part of the funnel that's generating the demand that closes later.

Micro conversions exist to cover that gap. Counting how many demo requests a channel generates, how many leads download a case study, or how many people repeatedly visit the pricing page gives an early business signal, available weeks or months before the macro conversion materializes. They don't replace the sale as the final metric, but they let you judge whether a channel is working long before the full cycle plays out.

The piece usually missing to close the loop is connecting those micro conversions to the real CRM outcome: which leads, from which channel, ended up signing. Without that data match, any conclusion about which campaign "performs better" rests only on the part of the funnel GA4 can see on its own, which in B2B is usually the less important half.

Anyone who's spent years in web analytics still says "goal" out of habit, the Universal Analytics (UA) term, the version of Google Analytics that preceded GA4. According to Google's official documentation, standard UA properties stopped processing new data on July 1, 2023, and access to already-processed data in the interface and API ended on July 1, 2024 (verified source, see Sources). In UA, a goal was configured at the view level and could be of type destination, duration, pages/session, or event-based, with a limited number of active goals per view.

GA4 replaced that model with one built entirely on events. There are no separate "views" or goals configured on the side: every interaction is captured as an event, and any existing event can become what Google's current documentation calls a key event. The process, per official help documentation (see Sources), runs through Admin, under Events: create or find the event, toggle "Mark as key event," then set that event's default value and counting method, which can count every instance of the event or only once per session.

Managing these settings requires, according to the same documentation, a Marketer role or above within the property; a Viewer role only allows viewing reports, not marking new events as key events. That permissions detail explains a fair amount of confusion in teams where marketing wants to adjust key event configuration but only has read access.

One detail worth checking case by case: the exact date Google internally renamed "conversion" to "key event" inside the GA4 interface, along with the maximum number of key events allowed per property, don't appear with a specific figure or date in the official documentation checked for this article. These are details widely cited across industry forums and blogs that we couldn't confirm against Google's primary source in this write-up, so we're flagging them as unverified and recommending you check them directly in your own property's Admin panel before planning a migration.

What is documented: "conversion" hasn't disappeared entirely. It remains the term Google Ads uses for the part of the integration that connects GA4's key events to campaign reporting, per Google's own help pages. The word coexists with "key event" depending on whether you're inside Analytics or inside Ads.

• Define what your actual business macro conversion is before touching anything in GA4's configuration; the tool won't decide that for you.
• Mark only micro conversions with real business value as key events, not every click that happens to be easy to measure; too many key events dilute the main report.
• Pick the right counting method for each key event (once per event or once per session), depending on whether you care about every purchase or just whether the user bought at all in that session.
• On long B2B cycles, connect the CRM outcome back to the original session source through offline conversion imports, so the channel that produced the lead gets credit when the contract closes months later.
• Check conversion rate by segment (channel, device, campaign) instead of only the aggregate number, which hides large differences between traffic sources.
• Test page variants with A/B testing before assuming the problem is traffic rather than conversion, and check the call to action first when the rate drops.
• Periodically audit which events are still marked as key: an event that's lost its business meaning keeps counting anyway and quietly skews reports.

• Marking every minor interaction, like a menu click, as a key event just because it's easy to measure, which inflates the "key events" count without adding real signal.
• Comparing the conversion rate of a macro conversion against a micro conversion as if they were the same metric, when they reflect very different user behavior.
• Writing off a B2B campaign as a failure because it hasn't produced macro conversions in the first few weeks, without accounting for a real cycle that can run months.
• Trying to rebuild the exact same goal structure GA4 had in Universal Analytics, with the same limits and types, when the event model works differently from the ground up.
• Not checking the permission role needed in GA4 before hunting for the option to mark key events with a Viewer-only account.
• Trusting an automatic event like "purchase" without first verifying it's firing correctly and counting accurately, before basing budget decisions on that data.