Case study · Banking · 17 November 2026

KRITIQ ranked five creator videos. The market ranked them the same way.

A banking brand put five creator videos through KRITIQ before a R200k TikTok campaign. KRITIQ ranked them on weighted risk and flagged one as Do Not Deploy because the brand only appeared at 1:05. We then compared KRITIQ's order with how many paid views actually reached the brand.

BankingTikTokR200k mediaMeasured · platform data + brand lift
KRITIQ score vs views reaching the brand
A74.2%0:004,730
B68.2%0:342,960
C62.2%0:141,580
D60.4%0:332,240
EDo Not Deploy52.1%1:05500

Top two and bottom video ranked exactly. Only C and D swapped places.

0.9rank correlation between KRITIQ's score and real brand exposure
9×more brand views per rand from KRITIQ's top-rated video than from the Do Not Deploy video
99.3%of plays on the Do Not Deploy video ended before the brand appeared. R21k of reach spend still ran behind it.
The brand lift study

Where viewers reached the brand, recall moved. Where they didn't, it stayed flat.

Overall ad recall, familiarity and recommendation showed no significant lift, which is what KRITIQ's late-branding flags predicted. The one place recall moved was the Video Views campaign, where people watched longer and three to four times more of them reached the brand.

Ad recall · Video Views campaign
Control18.8%
Exposed23.8%
+5.0 pts
Ad recall · Reach & Frequency campaign
Control23.8%
Exposed25.0%
Flat

How this was measured. Each video's brand-reveal time comes from its KRITIQ report, recorded before launch. Views reaching that moment are estimated from TikTok's 25/50/75/100% view-through data across both campaign objectives (Video A measured at its first brand mention, 0:10). The 0.9 is a Spearman rank correlation across five videos, which is a small sample, and reaching the brand is a proxy for awareness. The brand lift study was a TikTok Brand Lift Study (Sept 2026, about 900 people per group); the Video Views lift is significant at 90% confidence on about 350 people per group. We read this as consistent with KRITIQ's flags, not as proof that KRITIQ predicted the lift result.

FAQs

How we validate KRITIQ

Can't find your answer? Email hello@kritiq.app.

Are these real campaigns?

Yes. Each case study is a real campaign run by a KRITIQ client. Brand names are removed. The page says for each number whether it was measured (platform data or a brand lift study) or modelled from case study data.

When were the KRITIQ scores recorded?

Before launch. The creatives were audited in KRITIQ before any media was spent, and the scores were then compared with what happened in market.

What does a 0.9 correlation mean?

KRITIQ ranked five creator videos from best to worst before launch. We then ranked them by how many paid views actually reached the brand. A rank correlation of 0.9 means the two orders were almost identical: the top two and the bottom video matched exactly, and only two videos in the middle swapped places.

How is the ROI calculated?

For one R100k campaign with six creatives, we add the media protected by fixing weak creatives (a conservative 30 to 40% of the budget; the fintech case measured 62%) and the team time saved (valued at R1,000 an hour), then compare that with KRITIQ's entry plan fee. It is a projection: change the budget and hourly rate to match your team.

Can KRITIQ Studio run this on our campaigns?

Yes. Every KRITIQ Studio engagement starts with a diagnostic: we run your creatives through KRITIQ before launch, then link each one to its live results in KRITIQ Data so you can see the same validation for your own brand.

Stop guessing.
Start knowing.

Try KRITIQ free for 14 days. Audit your next five creatives, your website and the questions your customers ask AI.