How to measure the impact of growth channels

Every ad platform gives you a report saying it delivered excellent results. Add up the sales each platform claims, and the total may be larger than your actual sales. That isn’t a rare exaggeration; it is the natural result of how each platform counts what it considers “its” result.

Measuring the impact of growth channels means answering a harder question than “how many results did the channel record?” It means asking: what would have happened if we hadn’t invested in this channel? This article explains how to get closer to that answer in practice.

Credit isn’t impact

When a customer buys after seeing an ad, the platform credits the ad with the sale. But perhaps that customer would have bought anyway: they knew the brand, searched for it by name, and saw the ad on the way. The ad “got the credit,” but it didn’t create the result.

That is why it helps to separate two things:

• Attribution: how credit is shared among the touchpoints a customer passed through before buying.
• Incrementality: the sales that would not have happened without the channel.

Attribution is useful for day-to-day comparison and campaign optimization, but a big budget decision needs as much understanding of incrementality as you can get.

Why platform numbers differ from yours

• Different attribution windows: one platform counts a purchase that happens within a set number of days of a click or view; another uses a different window.
• View-through credit: some platforms credit a purchase to an ad that was seen but not clicked.
• Overlap: the same customer saw ads on two platforms, and each counts them as its own.
• Tracking limits: privacy settings in browsers and devices hide part of the path, so platforms rely on modeling and estimates.

The practical conclusion: don’t add platform numbers together, and don’t treat any of them as the final truth. Make your main reference what was actually recorded in your own system: the store, your customer relationship management (CRM) system, or your sales records.

Three layers of measurement

Layer one: a single source of truth for sales

Before any analysis, agree on the number that represents real sales or qualified customers, and where it comes from. Then connect each order to its source wherever possible: consistent tracking parameters in links, a source field in forms, and one simple question at purchase or on the first call: “How did you hear about us?” That simple question sometimes reveals what the tools don’t.

Layer two: comparing channels on one standard

Use a single attribution model in an analytics tool that is independent of the platforms, such as GA4, to compare channels with each other. The model won’t be perfect, but it applies the same rule to everyone, which is enough to see trends and large differences.

Layer three: impact tests

Big budget questions need an experiment that reveals incremental impact. Practical options include:

• Pausing or reducing spend temporarily: cut spend on a channel for a defined period, and watch what happens to total sales — not just to that channel’s numbers.
• Geographic tests: run a campaign in some cities and pause it in similar ones, then compare. Large Saudi cities can differ in behavior and demand, so choose areas that are as similar as possible.
• Platform lift studies: some platforms split an audience into a group that sees the ad and one that doesn’t. Useful — while remembering that the party running the test also benefits from its outcome.

A hypothetical example: reading a pause test

Imagine a store spending on a retargeting campaign that the platform credits with around a quarter of sales. The team decides to cut its spend to a minimum for three weeks, in a period free of seasons and major promotions, while keeping everything else the same.

If total sales fall by roughly what the platform was claiming, the campaign creates real impact. If the total stays almost flat, most of what it claimed would have happened anyway, and part of its budget can be redirected. Either way, the result is approximate and needs careful interpretation, but it is closer to the truth than the platform report alone.

Channels with long sales cycles

In business-to-business (B2B) sales, weeks or months can pass between first contact and a signed contract, with several people involved in the decision. Measuring channels by forms submitted is misleading here: some channels bring many forms that never convert, others fewer but more serious opportunities. It is better to follow each source to later stages — did the request become a meeting? A proposal? A contract? That requires linking the source of each request to your CRM, which we explain in How to build a measurement framework.

Channels that are hard to measure by clicks

Some channels work slowly or indirectly: content, social presence, PR and search engine optimization (SEO). Their effect often shows in indicators such as searches for the brand name, direct visits, faster deal closing and better conversion in other channels.

Hard-to-measure doesn’t mean no impact — but it doesn’t justify continuing with no evidence either. Agree on reasonable leading indicators for each channel and watch their direction over a longer period.

What to look at in the regular review

• Total sales or qualified customers, and their sources as your system records them.
• Acquisition cost per channel on one standard, compared with customer value and margin.
• Customer quality from each channel: do they come back? Do they request many refunds? Do their requests turn into contracts?
• The results of any impact test during the period.
• The decision: what do we increase, what do we cut, and what do we test next?

To interpret ad numbers specifically, read How to read return on ad spend (ROAS).

Common mistakes

• Judging a channel on one week: channels need enough time to settle, especially after major changes.
• Comparing channels on different standards: one channel measured by clicks and another by sales can’t be compared.
• Ignoring customer quality: the cheapest channel for acquisition may bring the least valuable customers.
• Building measurement after spending: it should be built before scaling, not after. We explain how in How to build a measurement framework.

In summary

Measuring channel impact is a journey from rough accuracy to better accuracy, not one final answer. Start with a single source of truth for sales, use one standard for comparison, and test the impact of large channels before doubling their budget. Above all, make every review end in a decision. That link between measurement and decisions is what we build in the Data, Measurement & Automation layer.