Benchmarks

Alltold teamed up with Innovid to measure the most served ads each quarter, starting in Q1 2024. These analytics are aggregated and shared quarterly to give the industry a view of inclusion, and to provide advertisers with benchmarks to compare their own inclusion analytics to.

Latest quarter: Q3 2024

Gender expression

Top served ads of Q3

Distribution of ads by screen time allocated to people with feminine gender expressions

Distribution of ads by screen time allocated to people with masculine gender expressions

Age & Body size

Age: Top served ads of Q3

Body size: Top served ads of Q3

Skin tone

Top served ads of Q3

Skin tone x Gender expression

Sexual orientation

Straight
Gay or Lesbian

2894

People detected

280

(9.7%)

Shown with a sexual orientation

278

(98.9%)

Straight

2

(0.7%)

Gay or Lesbian

Visible disability

Visible disability
Diabetes
Partial or completely missing limbs
Deaf / hard of hearing
Down syndrome
Mobility disabilities
Dwarfism

2894

People detected

29

(1%)

Shown with a visible disability

13

Diabetes

(All of these are pharma or med tech ads)

7

Partial or completely missing limbs

5

Deaf / hard of hearing

2

Down syndrome

1

Mobility disabilities

1

Dwarfism

Quarterly trends

Gender expression

Average % screen time

Age & Body size

Average % screen time

Average % screen time

Skin tone

Average % screen time

Average % screen time

Sexual orientation

Total people detected shown with a sexual orientation

% of people detected shown with a sexual orientation

Visible disability

Total people detected shown with a visible disability

Total people detected shown with a visible disability, excluding diabetes

Methodology

I. Ad selection:

  • Inclusion criteria: top 500 ads, ranked by # of impressions per quarter
  • There is some duplication of ads due to the same ad being run across different platforms with different titles

II. Ad evaluation:

  • We first ingest the ads into our system
  • We use a commercially available person detection algorithm to detect the people in the ads
  • We annotate each person detected using our identity attribute scales, built responsibly in collaboration with experts and advocacy groups like GLAAD, RespectAbility, and University professors
  • We use those annotations to build the analytics you see in these reports

III. Vertical analysis:

  • We exclude vertical analysis for verticals with less than 10 ads analyzed, or less than 3 advertisers represented. Those ads and advertisers are still included in the overall analysis

Download the full report




Submit

Thanks!

Your submission has been received and you will be redirected to the report shortly.
Oops! Something went wrong while submitting the form.