Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

YouTube Marketing › YouTube Analytics

Measuring YouTube's Contribution to the Business

How to show what an organic YouTube channel does for the business: tagged visits and key events, untagged influence, self-reported answers and branded search.
Guideadvanced11 min readUpdated

Why YouTube's contribution is hard to see

A paid campaign is bought to produce clicks and is measured on them. An organic YouTube channel is usually built to earn attention, trust and familiarity, and most of its effect happens on YouTube, where web analytics cannot see. A viewer who watches a tutorial and buys a month later through a search for the business's name leaves no trace of the video in the purchase record.

Three gaps compound. Most viewers never click a link, so their influence is invisible from the start. Of those who do click, many arrive without evidence of where they came from: the traffic sources guide notes that untagged clicks from YouTube arrive as referrals from the YouTube domain on the web and often as direct traffic from the apps, where no referrer is passed. And YouTube Analytics reports what happened on YouTube, not what happened afterwards, so the two sides measure different things in different units and never reconcile exactly; the guide to why platforms and analytics disagree explains the general causes.

The result is predictable. Reported on last-click website data alone, a channel looks small; reported on views, it looks large and says nothing about the business. Views, subscribers and watch time are the channel's own measures and the right ones for running it, as the YouTube Analytics hub sets out, but presented to the business as results they become vanity metrics. The job is to assemble evidence between those two extremes.

The chain from view to outcome

The evidence is easier to assemble when the chain it describes is written down first. A KPI tree breaks an outcome into the metrics that produce it, level by level, and the KPI tree framework shows how to read it for the lever that moved. For a YouTube channel whose job includes sending people to the business, the chain has recognisable links.

Link in the chainWhere it is measuredHow much of it can be seen
Viewers reached and watchingYouTube Analytics: views, unique viewers, watch timeFully, on YouTube's own terms
Viewers returningYouTube Analytics: returning viewers, browse and notification viewsFully, as an audience measure
Clicks to the websiteYouTube Analytics for end-screen element and card clicks; web analytics for the visits that arrivePartly; the visit side is seen only in web analytics, and is identified as YouTube reliably only when tagged
Visits from YouTubeWeb analytics: sessions with YouTube campaign tags or a YouTube referrerPartly; app clicks often arrive as direct
Key events from those visitsWeb analytics: enquiries, sign-ups, purchasesPartly; only where the visit was identified
Later outcomes influenced by the channelCustomer answers, branded search, direct visitsIndirectly, as supporting evidence

The table makes the measurement problem visible: the links nearest the business are the ones measured least directly. The measurement plan tutorial turns a tree like this into a document that names the events, the dimensions and the owner of each number, and the measurement plan framework is the shape to follow. For the YouTube channel, the plan records which outcomes count, how YouTube visits are identified and which supporting signs will be read.

Tagged visits and key events

The firmest evidence is the visits the channel can be shown to have sent and what those visitors did next. It depends on tagging every link the channel controls. Google Analytics' URL builder page, as the naming conventions and UTMs guide describes it, collects campaign parameters on links you control as campaign data. The traffic sources guide sets the YouTube rule: one source value for the platform, a medium that says the link was organic rather than paid, a campaign that names the video or series, and a content value for the placement. With that in place, every description link, pinned comment, channel page link and end-screen link to a website where the channel is eligible produces sessions that can be attributed to a video and a placement. The guide to sending YouTube viewers to your website covers which videos should carry links and where.

The outcomes those sessions lead to are read through key events. The key events and conversions guide describes a Key Event as an ordinary event the analytics property has been told to treat as an outcome, and advises marking few, meaningful ones. For a YouTube report, use the same key events the business already reports for every other channel, such as an enquiry, a quote request, a sign-up or a purchase, so the channel is judged on the same outcomes as the rest.

Read tagged outcomes in two ways. The last-click view counts outcomes in sessions that came from YouTube. The assisted view, as the guide to measuring content performance explains for content in general, counts journeys in which a YouTube session appeared before the converting session, whether or not it was the last touch. For a channel whose viewers usually return later by another route, assisted conversions are the more honest reading, with the caution that every Attribution figure is a modelled estimate under the model the tool applies.

Tagged evidence is the lower bound. It is firm because each visit is identified, and it is incomplete for the same reason.

Influence that arrives untagged

Most of a channel's influence never carries a tag. Viewers who watch on the YouTube app and tap a link may arrive as direct. Viewers who remember the business and search for it later arrive as organic or paid search on the business's name. Viewers who share a video in a messaging group send people who arrive with no trace. And some viewers buy somewhere the website cannot see: by telephone, in person, through a marketplace or a reseller.

The diagnosis for direct traffic that is too high explains how direct traffic is formed and how to locate the leaks, and its prevention habit is the one that matters here: tag every link you control. Once that is done, what remains untagged is genuinely beyond reach, and the channel's contribution to it can only be estimated from other evidence. The next two sections cover the two kinds of evidence most channels can gather.

Do not try to reclaim untagged traffic by assumption. Crediting every rise in direct visits to the channel because a video did well that week is the mirror image of ignoring the channel: both replace evidence with a story.

Asking customers where they heard of you

The cheapest way to see behind untagged influence is to ask. The self-reported attribution guide sets out how: an open question about how the person heard of the business, asked at a consistent point such as the enquiry form or the first call, coded into a stable scheme and compared with the tracked source on the same record. It lists audio and video whose call to action is a name to remember among the touches tracked attribution cannot see, which describes most of what a YouTube channel does.

For a YouTube channel, three refinements help. Make sure the coding scheme has a YouTube code, so answers that name the channel, a video or the presenter are grouped together. Compare the answers with the tracked source: a customer who names YouTube but arrived through branded search is exactly the influence the tagged view misses. And read the trend rather than the level, because the guide's list of biases applies here too, including memorability, which favours a face and a voice, and naming confusion, which can blur organic videos with advertising seen on YouTube.

The self-reported guide is clear about what the answers are: evidence about where attention came from, not proof of what a channel caused. Treat a rising share of answers naming YouTube, after the channel's activity increased, as corroboration, and never as a measurement of revenue caused.

Branded search and direct visits as supporting signs

When a channel is working, more people know the business's name. Two signals tend to move with that: branded search for the name and direct visits to the site. The self-reported guide suggests comparing self-reported trends with independent signals that move for the same reason, such as branded search volume after a campaign, and treating agreement between them as corroboration. The same applies to a YouTube channel: a rise in branded search over the months a series ran, in step with more customers naming YouTube, is a stronger case than either alone.

Both signals are crowded. Branded search and direct visits rise for many reasons, including other marketing, word of mouth, press, seasonality and growth in the business as a whole, so neither can be credited to the channel on its own. Read them over the same periods as the channel's activity, against the business's own history, and note what else changed. Where the numbers are small, the guide to reading small numbers advises looking only for large, sustained movements.

The route to causal evidence is an experiment. The incrementality testing guide describes withholding a channel from a comparable group and measuring what disappears, which is practical for paid media. An organic channel rarely allows a clean holdout: published videos stay findable by anyone, and pausing uploads leaves the library working. Say so in the report, rather than implying a precision the evidence does not have. The paid side of YouTube is measured differently, in Google Ads, and the YouTube Analytics hub points to where it belongs.

Reporting the channel to the business

The report a business needs from its YouTube channel is short, structured and honest about its limits. The KPI tree gives it a shape; the measurement plan says where each number comes from. The guide to social media reporting argues for reports that compare a channel against its own history and end in actions, and the same discipline applies here.

Layer of evidenceWhat it showsHow firm it is
Channel healthViews, watch time, returning viewers, subscribers gained, against the channel's own baselineFirm about the channel; says nothing about the business on its own
Tagged outcomesSessions from tagged YouTube links and the key events that followed, last-click and assistedFirm, and a lower bound
Customer answersThe share of enquiries or customers who name YouTube, and how many of them arrived untaggedDirectional; shaped by memory and wording
Supporting signsBranded search and direct visits over the periods the channel was activeWeak alone; useful when it agrees with the layers above
CostsProduction time and spend behind the videos in the periodFirm, if recorded

State the contribution as a range. The lower end is the tagged outcomes. The upper end adds outcomes where the customer named YouTube but arrived by another route, and a note on what the supporting signs suggest. Name the gaps: app traffic arriving as direct, sales the site cannot see, the biases in self-reported answers, and the lack of a holdout. The guide to reading YouTube audience reports covers the channel-health layer, which explains why the business layers moved.

End with a decision. A range that has held for several periods is enough to decide whether to keep the channel's investment, change its formats or move effort elsewhere, which is all the business needs from the report. The academy exercise on building evidence of what TikTok and YouTube add for a fishing tackle retailer practises writing the evidence and the report an organic channel can honestly support, for an owner who credits it with every rise in sales.

A loft-conversion builder builds its YouTube evidence (illustrative)

Illustrative example — not a real result

Hettravon Lofts reports its YouTube channel for six months

Hettravon Lofts, an invented loft-conversion builder, runs a project walk-through series: one video a fortnight showing a finished conversion with the homeowner's permission, 13 videos over six months. Every description and pinned comment carries a tagged link to the quote request page. Over the six months, tagged YouTube sessions total 1,240 and lead to 26 quote requests on a last-click basis; YouTube sessions appear earlier in the journeys of a further 19 quote requests that converted through search or direct. The quote form asks how the person heard of the business, and 58 of the period's 310 quote requests name YouTube, a video or the presenter, against 12 in the six months before the series. Of those 58, 31 arrived through branded search or direct rather than a tagged link. Branded search impressions for the business's name rise by about two-fifths on the same months a year earlier, while the builder's other marketing was unchanged. The owner reports the channel as contributing between 26 and 57 quote requests in the period: 26 firmly tagged, 45 including assisted journeys, and up to 57 when the 31 customers who named YouTube but arrived untagged are added, with branded search moving in the same direction. The report lists the gaps: app clicks arriving as direct, the biases in self-reported answers, and no holdout. Against the filming and editing time recorded for the series, the owner decides to continue the series for another six months and to add the presenter's name to the spoken ask in each video, so customers who remember the presenter can find the business.

The figures belong to the example. What transfers is the shape of the report: a firm lower bound from tagged outcomes, an upper bound from customer answers, supporting signs that agree or do not, the gaps named, and a decision at the end.

Common mistakes

  • Reporting views as results. Views and subscribers describe the channel; the business needs outcomes, even approximate ones.
  • Reporting last-click alone. Tagged last-click outcomes are a lower bound, not the channel's contribution.
  • Leaving links untagged. Every untagged link sends YouTube's visits into referral or direct, where they cannot be credited.
  • Using different outcomes for YouTube. Judge the channel on the key events the business reports for every other channel.
  • Treating customer answers as proof. They show where attention came from, not what the channel caused.
  • Crediting every rise in branded search to the channel. Other marketing and the business's own growth move it too; read it as support, not as a measurement.
  • Claiming an incrementality result the channel cannot produce. An organic library rarely allows a clean holdout; say so.
  • Presenting a single number. A range with its limits survives scrutiny; a single figure invites a dispute.

Frequently asked questions

Can I see YouTube conversions in Google Analytics?

Only for visits that Google Analytics can identify as coming from YouTube. Links you control in descriptions, pinned comments, the channel page and end screens can carry campaign parameters, which the analytics tool collects as campaign data, so those visits and the key events that follow them can be reported. Untagged clicks often arrive as referrals or as direct traffic from the apps, and viewers who remember the business and come back later through search are not credited to YouTube at all.

How do I calculate the return on an organic YouTube channel?

Start with the costs, which are usually the time and production spend behind the videos, and set them against the outcomes the channel can be shown to have contributed to, stated as a range rather than a single figure. The lower end is what tagged visits led to; the upper end adds the evidence from customers who name YouTube when asked and from supporting signs such as branded search. A single return figure built on last-click data alone will understate the channel, and one built on every signal at full weight will overstate it.

Can I run an incrementality test on an organic YouTube channel?

Rarely in a clean form. Incrementality tests withhold a channel from a comparable group, which is practical for paid media where delivery can be switched off for a region or audience, but published videos stay visible to anyone who searches for them. Pausing uploads for a period is possible, but the existing library keeps working and other changes happen at the same time. For most organic channels the honest evidence is a combination of tagged outcomes, customer answers and supporting signs, reported with its limits.

Terms used on this page

Frameworks

Where this page is used

Part of these learning paths

Also related

Sources

  1. YouTube Help — YouTube Help
  2. [GA4] URL builders: Collect campaign data with custom URLs — Google Analytics Help

Related