How often should a manager review artist data?

Weekly, for about twenty minutes per project, with a longer review at month end. Weekly cadence exists to catch stalls and broken data feeds while there is still time to act, which a monthly review is too late for.

What should a weekly music data review cover?

Four things in a fixed order: what changed against the project's own recent range, what is broken or has gone flat, what has stalled such as ticket pacing on an upcoming show, and what needs a decision this week. Then two written sentences per project recording what happened and what you decided.

How do you review data across a roster of eight projects?

Stagger the reviews across the week rather than doing all of them on Monday, and let tooling handle triage: a digest naming which projects moved outside their own range and which have a source that stopped syncing turns eight full reviews into two.

A weekly data review that takes twenty minutes

A repeatable Monday routine for managers running several projects: what to look at, in what order, what to write down, and what to deliberately ignore until month end.

Danny Angove

Co-founder, Backline · 19 June 2026 · 4 min read

In short

Twenty minutes a week per project beats two hours a month, because most of the value in this data is catching things while there is still time to act.
Go in a fixed order: what changed, what is broken, what is stalling, what needs a decision. Fixed order stops you following the most dramatic number.
Write two sentences per project every week. The archive of those sentences becomes the most useful document you own.
Deliberately ignore anything cumulative, anything with fewer than fourteen days of history, and any single metric that moved alone.
The output of a review is either a decision, a diary entry or nothing. If it is nothing, say so and move on.

Managers do not have a data problem, they have a time problem. The dashboard is open for four minutes between calls, and whichever number is largest wins the attention. That is how a rounding artefact ends up in a label email.

The fix is a fixed order and a hard time limit. Same sequence every week, in the same place, so you are comparing like with like and the biggest number does not get to set the agenda.

The order, and why it is that order

Illustrative

1What changed, 4 minutesSeven day window, movement only, compared against this project's own recent range. Never totals.
2What is broken, 4 minutesEvery source still connected, nothing gone flat, everything you expected has appeared.
3What is stalling, 5 minutesTicket pacing per show, campaigns that stopped working, markets that plateaued. The least dramatic and most useful step.
4What needs a decision, 5 minutesOnly for items the first three steps raised. Answer is do it, diarise it, or nothing.
5Write two sentences, 2 minutesWhat happened, what you decided. The archive is worth more than any single week's review.

Fixed order matters more than the exact minutes. Without it, attention goes to the largest number on the screen, which is frequently an artefact.

Minutes 1 to 4: what changed

Set the window to seven days, and look only at movement, never at totals. Streams gained, follower change, listener change, tickets sold this week, links clicked.

You are looking for anything outside the normal range for this project. Not outside a benchmark for other artists, outside its own recent range. A project that gains 12,000 streams a week is not having a good week at 13,000.

If nothing is outside range, write "quiet week" and move on. Most weeks are quiet weeks, and recording that honestly is what makes an unusual week visible later.

Minutes 5 to 8: what is broken

Faster than it sounds, and the step people skip.

Is every source still connected? Tokens expire, ad account grants get revoked, a platform changes something.
Has any chart gone flat rather than down? Flat is the signature of a broken feed. Down is a real decline.
Did anything you expected to appear not appear? A show you added, a campaign you launched, a release that went live.

A week of missing data found on the day it happens costs you nothing. Found at month end it costs you the month's reporting.

Minutes 9 to 13: what is stalling

The most valuable four minutes, and the least dramatic.

Ticket pacing per upcoming show: has anything stopped moving? A show four weeks out that has not sold in eight days needs a response now, not in the last fortnight when the options are discounting and worry.

Then the same question on the online side. A campaign that was working and has stopped. A market that was climbing and has plateaued. Search demand that has drifted down since the release.

Stalls do not announce themselves. They look like nothing happening, which is why a routine finds them and a dashboard glance does not.

Minutes 14 to 18: what needs a decision

Only now do you look at anything that requires judgement, and only for the things surfaced by the previous three steps.

For each one, the question is what changes as a result. There are three legitimate answers: do a specific thing this week, put it in the diary for a date, or nothing. "Keep an eye on it" is not an answer, it is a way of avoiding a decision, and it guarantees the same item appears next week.

Minutes 19 and 20: write two sentences

Per project, every week. First sentence: what happened. Second sentence: what you decided.

Streams flat, tickets in Manchester stalled at 61 percent with five weeks to go, Berlin listeners up a third after the radio play. Added a Manchester local push for Thursday, asked the agent about a Berlin date in the autumn.

That takes ninety seconds and it compounds. Six months of those sentences answers questions no dashboard can: when did this market start moving, what did we try in March, how long did the last placement take to decay. It is also the raw material for a label update, and it means the monthly report is a summary rather than an archaeology project.

What to ignore until month end

Weekly, monthly, or never

Illustrative

	Weekly	Month end	Never as a KPI
Streams gained in window	Yes	Yes	No
Total streams all time	No	As a header figure	Yes
Follower change	Yes	Yes	No
Save rate and conversion ratios	If available	Yes	No
Ticket pacing per show	Yes	Yes	No
Sources with under 14 days of history	No	Only with a label	No
Ad impressions	No	No	Yes
Press mentions, last 7 days	Yes	Yes	No

The middle column is not a lesser version of the first. Conversion ratios and comparisons need a longer baseline to be worth reading at all.

Cumulative totals are the main one. They are the same number every week by definition, so looking at them weekly trains you to feel progress where there is none.

Anything with less than a fortnight of history is the second. New sources, new shows, new campaigns: there is no baseline yet, and a percentage against a two day baseline is noise wearing a suit.

And any single metric that moved alone. A real change leaves marks in several places. One number moving on its own is more often mechanical than meaningful, which is the argument in when a spike is not a hit.

Running several projects

Twenty minutes each does not scale past about four projects on a Monday morning. Two adjustments make it work.

Stagger them. Two projects Monday, two Tuesday, the rest Wednesday. The review works on a weekly cadence, not a weekly deadline.

Let the tooling do step one and step two. A digest that tells you which projects had movement outside their own range, and which have a source that stopped syncing, turns "check eight projects" into "check the two that need me". That is the main thing automated summaries are actually good for: not the analysis, the triage.

Common questions

How often should a manager review artist data?: Weekly, for about twenty minutes per project, with a longer review at month end. Weekly cadence exists to catch stalls and broken data feeds while there is still time to act, which a monthly review is too late for.
What should a weekly music data review cover?: Four things in a fixed order: what changed against the project's own recent range, what is broken or has gone flat, what has stalled such as ticket pacing on an upcoming show, and what needs a decision this week. Then two written sentences per project recording what happened and what you decided.
How do you review data across a roster of eight projects?: Stagger the reviews across the week rather than doing all of them on Monday, and let tooling handle triage: a digest naming which projects moved outside their own range and which have a source that stopped syncing turns eight full reviews into two.

Sources

1Spotify for Artists, Reviewed August 2026. Data in Spotify for Artists
2Pollstar, December 2025. 2025 Year End Business Analysis

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

The twelve numbers worth knowing about every project

A short list of music industry KPIs that survive contact with a real week, why each one earns its place, and the popular metrics that belong in a footnote instead.

Danny Angove · 3 min read

Analytics foundations

When a spike is not a hit

Six things that produce a chart spike with no underlying audience gain, how to tell them apart in an afternoon, and what to check before anyone reforecasts a campaign.