Influur says follower count had no correlation with views per dollar across the creators it has paid, challenging its use as a basis for allocating music marketing budgets. In its analysis of label campaigns, creator selection explained 62% of the variation in cost per view.
The company, which has run thousands of creator campaigns across music, beauty and consumer brands, reviewed every label music campaign it conducted over the past two years, including daily viewing histories for their creator posts.
What predicted campaign efficiency
Influur compared the most efficient 10% of label campaigns by cost per view with the least efficient 10%. It reported no distinguishing differences in their briefs, despite a cost-per-view gap exceeding 20 times.
Among creators with more than one million followers, the most efficient 10% generated views at about $1 per thousand, while the least efficient 10% cost more than $20 per thousand. Creators with 100,000 to 500,000 followers showed a 15-fold range.
Creator tier, genre, market and budget together accounted for the variation that creator selection did not explain. Across all creators the company has paid, it reported zero correlation between follower count and views per dollar.
Views divided by followers proved more useful: creators in the highest quarter for that ratio delivered reach at approximately one-fifth the cost of those in the lowest quarter. Influur said the relationship held across every creator tier.
Posts costing less than $300 were the least efficient price band in the dataset. The most efficient band was $1,200 to $2,000 per post.
Selection and budget allocation
Influur assesses sustained performance over four weeks rather than relying on a single viral post. It also examines the sounds creators used without payment during the preceding 20 days to assess genre fit, saying most creators regularly use two or three genres.
The company recommends giving creators the song, the relevant musical passage and the budget, while leaving execution to them. In one campaign initially aimed at lifestyle creators within a single category, mothers filming with their children generated the most views on the sound, prompting a budget shift toward those creators.
Viewing activity was concentrated among a small share of posts and campaigns:
- 1% of tracked posts generated 41% of all views.
- 10% of posts generated 75% of views.
- The top 10% of campaigns produced half of all views.
The most efficient campaigns in Influur’s dataset used six to ten creators.
When and what to measure
Across tracked posts, views accumulated on the following timeline:
- Day one: 16% of final views.
- Day two: 74%.
- Day seven: 91%.
- Day 14: 97%.
- Day 21: viewing activity was complete.
Influur recommends assessing campaign results on day 14 and using day-two performance to inform budget changes.
At campaign level, engagement rate had no correlation with views per dollar. At post level, the correlation was negative. Influur tested four definitions of virality across its post history; only one post met all four, and just 7.5% of posts classified as viral by views also qualified by engagement.
The company also recommends tracking streams, saves and Shazams during the two weeks after posting, alongside unpaid content using the audio.
Campaign example and new product
Earlier this year, Influur worked on Temper City’s debut single, Self Aware, released by independent label Thirty Knots. The campaign used 17 creators, none described as major names, with a $5,100 budget, generating 30.7 million organic views at a reported cost per thousand impressions (CPM) of $0.17.
The song subsequently entered the Billboard Hot 100, reached No. 1 on Spotify’s Global Viral chart and surpassed 150 million Spotify streams.
Influur says its Pulse product automates creator selection using the signals outlined in its analysis. Users upload a track and set a budget; the agent generates a creator list, identifies creators showing strong current performance and provides budget allocation guidance by day two. The company says the process takes four minutes, compared with 48 hours.
