A measurement plan
One document naming every event, what it means, which of them count as a conversion, and who agreed to that.
If you want long-term retainers Partnership and pricing
Google Analytics, Tag Manager and the events underneath, set up so the report and the books can be reconciled. Here that means accounting for the sale that finishes in a chat thread rather than in a checkout.
For teams already spending on marketing who cannot say which spend produced which sale. On a site with almost no traffic the honest next job is getting visitors, not counting them more accurately.
One document naming every event, what it means, which of them count as a conversion, and who agreed to that.
Containers, events and conversions configured in Google Analytics 4, with consent wired into the cookie notice rather than bolted on afterwards.
Because we work on the templates, the important events come from the code that renders them, so a design change cannot silently break a trigger that was matching a styling class.
Sales confirmed in your own system sent back to Google and Meta, so the platforms optimise against orders you were actually paid for.
Form submissions, chat conversations and phone calls tied back to the campaign, search term or page that produced them.
A short dashboard of the numbers that change decisions, with the definition of each written next to it.
Plenty of Philippine sales never touch a checkout page. The enquiry arrives in a Messenger thread, the payment is a GCash or Maya transfer confirmed by screenshot, or a deposit into your bank account, and the order is agreed in chat. Standard analytics records a page view and nothing else, so the report and the books disagree and in the end nobody trusts either. Consent is the other thing to get right, because what you collect needs a privacy notice people can find and a choice that genuinely changes what loads, rather than a banner that fires the same tags whichever button is pressed.
Usually neither, because they are counting different things. Analytics counts events in a browser inside its own attribution window; your sales system counts orders that were paid for. The job is not to force them to match, it is to know why they differ and which one answers which question. We write that difference down so it does not get relitigated every month.
The tag is enough for plenty of sites. Server-side earns its place when browsers and blockers are dropping a meaningful share of your events, when the sale completes somewhere the browser cannot see, or when you want conversions sent back to the ad platforms from your own records. Which of those applies to your setup is worth establishing before anyone fits it by default.
No, but it puts obligations on you. You need a privacy notice people can find, a lawful basis for what you collect, and a consent choice that actually changes behaviour when someone declines it. The National Privacy Commission also expects a personal information controller to have a named data protection officer. We configure analytics to respect the consent choice, not to record it and carry on regardless.
Yes, by different methods. Click to message ads and website chat can carry a campaign reference into the conversation, and phone calls can be attributed with a number that changes by source, provided calls are logged when they come in. What cannot be tracked is the walk-in who saw an ad last month and mentioned nothing, and that is better admitted than credited somewhere convenient.
Yes. The measurement plan stands on its own: it names the events, the parameters and what counts as a conversion. We can hand it over, review what gets built and test it afterwards. Most of the faults we find are not coding faults, they are two people meaning different things by the word lead.
Tell us what you are building and we will tell you honestly how we would approach it.