Customer story

Storyblocks picks creators with a score.

The stock-media company replaced gut-feel creator selection with a custom Trudy Score, built on two years of their own campaign data.

1.8×
the conversions from the same creator spend. At a flat subscription price, that's roughly 1.8× the revenue.
vs. Storyblocks' previous baseline · modelled on ~3,000 campaigns · 2024–2025
1.8×
Conversion uplift for high-Trudy-score creators vs. Storyblocks' previous baseline
~50%
Of past picks underperformed before Trudy
~3,000
Campaigns analysed · 2024–2025
2 yrs
Of their own first-party data
~50%
of the creators they paid underperformed
The challenge

About half the roster wasn't working.

Storyblocks sells stock video, audio and images by subscription, and pays thousands of creators on YouTube, Instagram and TikTok every month, with hundreds of partnerships live at any one time. At that scale, selection ran on gut feel and team relationships, with no number to defend a pick in a planning review.

A 50% miss rate, applied to thousands of partnerships a year, is real money going to underperforming creators.

What we did

We built them a score from two years of their own campaigns.

01

Share the history

Storyblocks handed over two years of past campaign data: who they ran, and how each one performed.

02

Train a Trudy Score

We built a custom model on that data to predict which creators would convert for Storyblocks.

03

Prove it on unseen creators

We held back part of the history and tested the score on creators the model had never seen. That test is where the 1.8× comes from.

Under the hood

How the score reads a creator.

A Trudy Score isn't built on follower counts. For every creator, Trudy analyses more than 160 research-backed signals of genuine influence, drawn from how a creator actually communicates.

01 · Language

Every word they say

The vocabulary, framing and narrative patterns a creator uses to carry an audience from interest to action.

02 · Delivery

How they say it

Tone of voice, pace and emphasis. The vocal cues that separate genuine conviction from a flat read.

03 · Expression

Their face on camera

Facial expression and on-screen presence that shape how trustworthy, warm and engaging a creator feels.

04 · Psychometrics

Personality & psychology

Profiles grounded in published behavioural research that proxy for trust and persuasion.

Those signals roll up into one score that tells you why a creator works.

The result

Higher Trudy Score, higher conversion.

Across a held-out test set, creators with a high Trudy score converted at 1.8× Storyblocks' previous baseline rate.

Same creator spend. 1.8× the conversions, and the new-subscriber revenue that comes with them.

1.8×
conversion uplift for high-Trudy-score creators vs. Storyblocks' previous baseline
Trudy's high-scorers 50 60 70 80 90 0 0.05% 0.10% 0.15% 0.20% 0.25% Trudy Score Conversion rate

Each dot is one creator from the held-out test set. Conversion rate climbs with Trudy Score. The strongest converters, shown in green, cluster in the high-score zone, where Storyblocks now concentrates spend.

The drivers

What set their best creators apart.

Which of those 160+ signals matter most is different for every brand. For Storyblocks, two stood out.

Driver 01 · Persuasion

Their top creators made the case.

Storyblocks' best creators were consistently persuasive. They leaned on comparison and causation, walking their audience through why a product was worth it.

Comparison

Language that weighs one option against another, giving the audience a frame of reference for the choice.

Causation

Language that connects cause to effect, explaining why something works.

Driver 02 · Values match

Their values lined up with Storyblocks'.

Top performers shared the values that matter most to Storyblocks: high on self-direction (freedom of thought over conformity) and benevolence (caring about others and community). When a creator's worldview overlaps with the brand's, the partnership reads as authentic.

Storyblocks values Top performers
Self direction Stimulation Hedonism Achievement Power Security Conformity Tradition Benevolence Universalism

Illustrative · Schwartz values profile

We work with thousands of creators. Once you figure out how to run a programme at that scale, the hard part becomes picking the right people. Trudy helped us put structure behind that instinct. We gave them our historical performance data, they built a model around it, and now we get a score before we commit budget.

Matt Reichard
Matt Reichard
Head of Influencer Marketing, Storyblocks
How they use it now

No budget moves before a score.

01

Score first

Every candidate is scored before any spend is signed off.

02

Drop or rework

Low scorers get dropped or reworked before they reach a campaign.

03

Back the winners

High scorers get the spend.

The team keeps their judgement. Now they have a number to back it up.

Score your roster

Score your roster like Storyblocks did.

Send us a year of your past campaigns. We'll score the creators and show you which ones the model would have backed. Then check that against what actually happened.