How do you measure whether music advertising worked?

Measure the first outcome you own end to end, which is clicks on your own landing page by platform and country, plus the click-through from that page. Then look at movement in streams, followers and ticket sales across the campaign window and for three weeks after spend stops. Report the second group as co-movement rather than as attributed results.

Why do platform-reported clicks not match my landing page visits?

Because the platform counts a click more generously than a page counts an arrival, and some people never finish loading the page. A gap is normal. Use your own landing page figure for anything downstream, and use the platform figure only for comparing adverts against each other.

Should ad spend across accounts and currencies be combined into one number?

Not without saying so. Ad accounts report in their own currency, and projects frequently have more than one account. A combined total needs both the account list and the exchange rate with its date attached, otherwise it describes a set of accounts and a rate nobody chose.

What to measure when you run ads for a release

Ad platforms report on themselves and report generously. The five numbers that tell you whether music advertising worked, and the three the platform shows you instead.

Danny Angove

Co-founder, Backline · 6 August 2026 · 3 min read

In short

An ad platform grades its own work. Cost per click and click-through rate describe the advert, not whether it did anything for the project.
Measure the first outcome you own end to end, which for most music campaigns is a click on your own landing page, by platform and country.
Then look for co-movement in the outcomes you cannot attribute: streams, followers, ticket sales. Report those as movement, never as attributed conversions.
Spend into markets that already respond. Widening into markets with no prior evidence is the most common way a release budget disappears.
Record what the money bought in audience terms, not in impressions. Followers per hundred spent survives comparison across campaigns.

Every ad platform gives you a dashboard that says the campaign went well. It is measuring reach, clicks and cost per click, all of which describe the advertising rather than the project.

The question you care about is different: did this money leave the project better off. Here is a way to answer it that does not require pretending attribution exists.

The three numbers the platform leads with

Impressions. Buy more and it goes up. It is a budget restatement.

Click-through rate. A creative quality signal, useful for comparing two adverts against each other and almost useless as a campaign outcome.

Cost per click. Genuinely useful for buying efficiency, and it tells you nothing about what happened after the click.

None of these is dishonest. They are simply about the advert, and there is nobody else to ask about the advert.

Four layers, decreasing certainty

Illustrative

1Measured by the platformImpressions, clicks, cost per click. Accurate, and about the advert rather than the project.
2Measured by you, end to endLanding page arrivals and click-through by platform and country. Your own data, fully attributable.
3Observed co-movementStreams, followers and ticket sales during and after the window. Real, and not attributable to a specific advert.
4InferenceAnything claiming a number of streams caused by a campaign. A model, not a measurement. Ask for the assumptions.

Certainty falls as you move down and usefulness rises. The skill is reporting the third layer honestly rather than dressing it up as the second.

The five to measure instead

1. Clicks on your own landing page, by platform and country

The first outcome you own end to end. If your adverts point at your own link page rather than straight at a streaming service, you can see how many people arrived, where from, and which platform they chose.

That last part is worth the setup on its own. Platform choice at the link page is the closest thing to a preference signal you get, and it is frequently not what the team assumed.

2. Click-through from the landing page

Of the people who arrived, how many went on to a service? A low rate points at the page rather than the advert: wrong artwork, slow load, missing platform, a broken destination on mobile.

This is the cheapest fix in music advertising and the least examined.

3. Movement in streams during and after the window

Not attributed. Movement. Take the streaming series, mark the campaign start and end, and look at whether the level changed and whether it held after spend stopped.

The bit after spend stops is the interesting part. A campaign whose effect ends the day the money does bought attention. A campaign with a higher baseline three weeks later bought an audience.

4. Follower and audience movement over the same window

Followers gained per hundred spent, roughly. It is a crude figure and it is comparable across campaigns, which is more than can be said for impressions.

If a campaign produced a lot of streams and almost no followers, you rented attention. That is sometimes the goal, for a chart week or a placement push, and it should be a decision rather than a surprise.

5. Ticket sales in the markets you targeted

For any project with live dates, this is the outcome that pays. Look at whether pacing changed in targeted cities against untargeted ones, which is a natural comparison the campaign itself created.

Asking what the spend actually bought

Illustrative

You asked

We spent on Meta and Google for the Aster Bloom single over three weeks. What did it buy, and did anything hold after it stopped?

Data readget_ads_performanceget_smart_link_statsget_streaming_statsget_audience_geography

Backline AI

Two ad accounts are connected for this project, so these figures cover both and I have listed them separately below.

Meta: 2,480 pounds, 41,900 clicks reported by the platform. Google: 1,150 pounds, 9,400 clicks. Each figure is in that account's own currency.

Your landing page recorded 18,600 arrivals in the same window, with 61 percent clicking through to a service. Spotify took 68 percent of those, Apple Music 17 percent.

Streams rose from a daily average of about 4,100 before the window to 9,700 during it. Three weeks after spend stopped, the daily average is 6,400, which is above the pre-campaign level.

Followers gained across the window were 3,120, which is roughly 86 new followers per hundred pounds spent across both platforms.

The gap between platform-reported clicks and landing page arrivals is normal and worth knowing about: the platform counts a click more generously than your page counts an arrival, so use your own figure for anything downstream.

Illustrative example using a fictional project. The assistant names the accounts covered rather than silently combining them, and reports the streaming change as movement rather than as attributed conversions.

Multi-account and currency traps

Two operational things that quietly corrupt ad reporting.

Multiple ad accounts. Projects often have more than one: a legacy account, a label's account, a new one. If a tool silently sums them, spend and results describe a set of accounts nobody chose. Any number should say which accounts it covers, and a request that is ambiguous should be answered with a question rather than a total.

Currencies. Ad accounts report in their own currency. Adding pounds to euros to dollars into one spend figure is wrong unless a rate and a date are attached, and if a rate is attached it should be visible.

Reporting it

Four lines and a chart:

Spend, by platform and account, in each account's own currency.
Landing page clicks and click-through, with the platform split.
Movement in streams and followers across the window, and the level three weeks after spend ended.
What you would do differently.

The chart is the campaign window marked on the streaming and follower series, so anybody reading can see the shape rather than take your word for it.

The comparison that matters

Not against a benchmark. Against your own last campaign, and against the counterfactual of not spending.

The second is harder and worth attempting: what was the project doing before spend started, and did it change beyond its normal range? Most of the time an honest answer includes "we cannot fully separate this from the placement that was running", and saying so is what makes the times you can say something definite worth listening to.

Common questions

How do you measure whether music advertising worked?: Measure the first outcome you own end to end, which is clicks on your own landing page by platform and country, plus the click-through from that page. Then look at movement in streams, followers and ticket sales across the campaign window and for three weeks after spend stops. Report the second group as co-movement rather than as attributed results.
Why do platform-reported clicks not match my landing page visits?: Because the platform counts a click more generously than a page counts an arrival, and some people never finish loading the page. A gap is normal. Use your own landing page figure for anything downstream, and use the platform figure only for comparing adverts against each other.
Should ad spend across accounts and currencies be combined into one number?: Not without saying so. Ad accounts report in their own currency, and projects frequently have more than one account. A combined total needs both the account list and the exchange rate with its date attached, otherwise it describes a set of accounts and a rate nobody chose.

Sources

1Meta, Reviewed August 2026. Graph API insights reference
2Google, Reviewed August 2026. Analytics Data API: report basics

Danny Angove

Co-founder, Backline

Danny Angove is a co-founder of Backline. He works on the music business side of the platform: what managers, labels and independent artists actually need to see, and which numbers are worth arguing about.

Backline does this for the projects you run

Streaming, audience, social, advertising, website, search, ticketing and press data in one dashboard per project, with an AI assistant that answers questions about your own connected data. Invite-only.

What Backline does

Analytics foundations

Correlation is the honest answer in music marketing

There is no passback linking a stream to the advert that caused it. What you can measure is whether signals move together, and how to report that without overclaiming.

Danny Starr · 3 min read

Marketing and campaigns

Smart links that earn their keep

A link page is the only step in a music campaign you fully control and fully measure. What to put on it, what to record, and the fan privacy line worth holding.

Danny Angove · 4 min read