New essays from the team — on casting, gifting, and what performance actually means
settr
Intelligence · the decision layer

Know which creator content deserves budget.

Settr Intelligence turns creator posts into commercial signals — then recommends what to boost, license, affiliate, test, or set aside. Not more reporting. Better decisions, before money is wasted.

/01 Turns Posts into signals

Every gifted asset scored on six dimensions — not vanity metrics.

/02 Recommends One move per asset

Boost, license, affiliate, test, or set aside — with the reason attached.

/03 Closes the loop The next list gets smarter

Paid and affiliate outcomes retrain the weights for the next round.

01 · What intelligence means

Not more reporting. Better decisions.

Most platforms report views, likes, and EMV after the campaign. Settr Intelligence decides the next move before money is wasted — and tells you why.

same data · different output
A Traditional reporting

"What happened?"

  • /01Views, impressions, reachnumber
  • /02Likes, comments, savesnumber
  • /03Earned media value (EMV)number
  • /04Engagement rate vs. follower countratio
  • /05Post-campaign recap deckpdf
output · numberscadence · monthly
B Settr Intelligence

"What should we do next?"

  • /01Boost from the creator handleaction
  • /02License for paid + organic re-useaction
  • /03Test as a brand-handle paid variantaction
  • /04Route to an affiliate cohortaction
  • /05Add similar creators to next month's giftingaction
output · decisionscadence · per asset · live
02 · The scoring system

Every post is scored in context.

A creator post is not good or bad in isolation. Settr compares it against six benchmarks — and the weights aren't static. They retrain every 14 days from your paid + affiliate ground-truth.

model · v2.4 · weights retrained 14d
/01
Creator contextDid this post outperform the creator's own normal content?
weight0.26score78
signal for boost · creator handle
/02
Niche benchmarkDid it outperform similar content in the broader category?
weight0.22score84
signal for paid test · usage rights
/03
Audience confidenceDoes the audience match the brand's target buyer?
weight0.18score67
signal for creator-context vs. broad test
/04
Affiliate intentDoes the creator show commercial buying behaviour?
weight0.16score71
signal for affiliate · community boost
/05
Paid readinessDoes the asset have the structure to work in paid social?
weight0.18score88
signal for paid social · A/B
/06
EditabilityCan the content become stronger paid variants?
weight0.10score62
signal for edit · request rights
weights · brand-tuned · retrained every 14 days model · v2.4 · sample brand build
03 · Decision logic

One post. Different moves.

The recommendation depends on which signals are strong. Three real-shaped examples — same scoring model, different signals dominant, different recommended move.

signal → recommendation → reason
example /01A
  • creator context+38%
  • niche benchmarkp57
  • audience matchhigh · premium
  • paid readinessp62
recommendation
Boost from creator handle.

The post works because of creator trust and a tight audience match. A brand-handle ad would strip the context that made it land.

example /02B
  • creator contextp51
  • niche benchmark+62%
  • audience matchbroad
  • paid readiness+22
recommendation
Request usage rights · brand-handle test.

The asset travels beyond the creator's own audience. The structure — hook, pacing, CTA — is paid-shaped. License it and test cold.

example /03C
  • creator context+28%
  • affiliate cadencehigh · LTK · codes
  • tracked linknone yet
  • audience matchwarm
recommendation
Move into affiliate · boost in community.

Her audience is already trained to act on recommendations. The post lacks a tracked link, but the cadence signal tells us where she'll perform.

04 · Audience confidence

Not all audience data is equal.

Some creators connect Graph API and ship stronger audience data. Others are modelled. Settr uses both — but weights them differently. Estimated data can guide. Verified data can decide.

source → confidence → weight
01
Content-level Graph APIhighest confidence

Audience data attached to the post itself, from the creator's connected Meta or TikTok permission. The strongest signal we can read.

confidenceverified
02
Creator-level Graph APIhigh confidence

Audience data attached to the account, not the specific post. Reliable for niche-fit calls, and the default when content-level isn't available.

confidenceverified
03
Settr-estimated audiencedirectional signal

Modelled from follower behaviour, comment patterns, and creator-graph proximity. Lower weight in the score — useful, not decisive.

confidenceestimated
04
No audience dataunavailable

No API connection and insufficient surrounding signal. Other dimensions can still score, but audience confidence is held at zero — never imputed.

confidence

Estimated data can guide decisions. Verified data can strengthen them. audience confidence is never imputed when missing

05 · Affiliate intelligence

Commercial behaviour matters before the sale.

Settr doesn't only wait for tracked sales. It also reads how a creator already behaves like a commercial creator — and uses both layers to route assets.

two layers · weighted together
A measured affiliate performance

What the tracked links say.

Click-throughs, sales, and per-click economics from tracked links. The clean signal — but only available after a creator has run a tracked partnership.

clicks · 30d1,284verified
orders423.3% CR
revenue€2,940AOV €70
EPC€2.29per click
RPM€18.20per mille
cohort tier02of 03
route to affiliate cohort B · commission tier 2 · re-seed next cycle
B affiliate cadence signal

How the creator already behaves.

Commercial cues read from recent activity — even before a tracked link has run. The signal arrives before the first conversion does.

LTK · weekly Amazon storefront ShopMy · 2 brands discount codes · 4 active retailer mentions · 6 / 30d "link in bio" cadence product reco language save-rate above baseline
route to affiliate · pre-tracked tier · ship sample + creator code · monitor 14d
06 · The paid feedback loop

Paid results close the loop.

When an asset is tested on Meta, the result flows back through the Marketing API and attaches to the original creator post. Spend stays in your accounts. Learning stays in your model.

post → growth → meta → insights → learning
feedback chain · asset A-2840 · cycle 14 spend window · 72h · closed
/01
Creator post

Organic asset enters the library. Scored, routed, awaiting activation.

@noor.beauté · REEL · 0:23
/02
Growth activation

Boost · paid · affiliate · license. Human approval, system runs the queue.

paid social · A/B test live
/03
Meta ad

Asset runs against your ad account. Settr never touches spend or pixels.

ad set · spend window 72h
/04
Ads Insights

Results pull back through the Marketing API into the asset's record.

auto-pull · 15-min sync
/05
Learning

Weights and routing rules updated. Same brand. Smarter next decision.

model · v2.4 → v2.5 (proposed)
spend€2,840
impressions186k
ctr3.7%
cpc€0.34
roas5.0×
purchases28

The paid result is linked to the original creator post — and feeds the next routing decision and next month's product-gifting list.

07 · Brand-specific learning

Each brand learns differently.

The six dimensions are the same for everyone. Which ones matter most is brand-specific — and the system learns it from your paid + affiliate results, not from a discovery call.

three archetypes · same model
premium · clinical

Brand Asmall premium audience

  • creator context0.26
  • audience confidence0.22
  • niche benchmark0.18
  • paid readiness0.12
  • affiliate intent0.12
  • editability0.10

Buyer cohort is small and hard to fake — audience confidence dominates. The system rewards creators whose followers are the buyers.

d2c · mass · paid-led

Brand Bscale on creative quality

  • paid readiness0.28
  • affiliate intent0.22
  • niche benchmark0.18
  • editability0.14
  • audience confidence0.10
  • creator context0.08

Spend scales when creative is paid-shaped. Paid readiness + affiliate intent dominate — the brand handle carries the ad.

affiliate-led

Brand Ccommission-driven

  • affiliate intent0.30
  • creator context0.22
  • niche benchmark0.16
  • audience confidence0.14
  • paid readiness0.10
  • editability0.08

Commission programmes reward repeat sellers. Affiliate cadence + creator trust dominate — the system goes deeper on creators who already convert.

See it run inside a programme How Settr works
Next

Find the content worth putting money behind.

A thirty-minute walkthrough on your own numbers. We show which creator posts to boost, license, affiliate, or test — and how every result makes the next creator list smarter.

You'll talk to Simon Fransson, Settr's founder.