Solutions
Built for the teams who carry the returns line.
ExactHue Core is a business-to-business engine. It is bought by brands, retailers and marketplaces, and it is judged on commercial outcomes rather than on demo quality.
Beauty brands
Complexion ranges live or die on shade accuracy. If you sell foundation, concealer, tinted skincare, cushion or lip online, mismatch returns are almost certainly your largest single return reason — and the one you have least visibility into.
Retailers & department stores
You carry dozens of brands whose shade nomenclature agrees on nothing. A measured colour layer gives you one comparable axis across the whole assortment — and lets a customer cross-shop shades rather than abandon the category.
Marketplaces
Third-party sellers upload swatches shot under whatever light was handy. Normalising those listings to measured colour reduces dispute volume and makes shade a filterable, trustworthy facet.
Where it is used
Four places measured colour changes the outcome.
Shade finder on the product page
The customer takes one capture. The engine qualifies the lighting, recovers reflectance, and returns a ranked shortlist with plain-language reasoning: which shade, how close, and why the runner-up differs. Where the capture cannot support a confident answer, the interface asks for a better one instead of guessing.
Typical placement: above the variant selector, replacing or augmenting the swatch grid.
Honest virtual try-on
The matched shade is composited back onto the capture under the customer's real lighting, with texture and specularity preserved. A lighting toggle shows the same shade in daylight and warm indoor light — the two conditions in which a customer will actually judge the purchase.
No smoothing, brightening or reshaping is applied at any point.
Cross-brand shade translation
“I wear 240N in another brand” becomes a solvable query once both ranges exist as measured colour. For retailers, this converts the most common in-store question into an online journey that ends in a basket rather than a bounce.
Requires both catalogues ingested; no capture needed from the customer.
Range planning & inventory signals
Aggregated, non-identifying match distributions show where demand actually sits across the tonal range — including demand your current range cannot serve. Gaps show up as clusters of customers whose best available match still exceeds a perceptible ΔE.
Delivered as a dashboard and a scheduled export for merchandising teams.
Measurement, not promises
How a pilot is evaluated.
We are a pre-seed company founded in 2025. We would rather show you a defensible method than quote someone else's case study.
Every engagement starts by establishing your own baseline, because industry averages are not your numbers. Improvement is then measured against that baseline on your own products, with shade-mismatch returns isolated from every other return reason.
Baseline
We take twelve months of return data and classify it by reason, shade and tonal band. This alone usually surfaces a pattern the team had not quantified.
Catalogue measurement
Your range is ingested as measured colour — from your own spectrophotometer data, or from physical samples measured in our laboratory.
Controlled rollout
The shade finder is enabled for a traffic share, holding the remainder as a control. Both cohorts are tracked on the same definitions.
Read-out
Mismatch return rate, single-shade order share, and per-band accuracy, reported against control. If the effect is not there, we say so.
What is included
Everything needed to run it properly.
The engine itself
Match, render and catalogue endpoints, in both a sandbox and production. The keys are scoped separately, so nobody can point a test script at live traffic by accident.
A widget you don't have to design
The shade finder picks up your own type and colour tokens, so it looks like your site rather than ours. Turn JavaScript off and it falls back to a plain shade grid.
Your range, measured
Send us physical samples and we'll measure them, or send readings you already hold and we'll check them. Either way you'll hear about out-of-gamut shades before launch, not after.
Accuracy you can interrogate
ΔE00 broken out by Fitzpatrick band and undertone family, refreshed with every model release. We don't publish a single blended number, because that's how the deep end of a range gets hidden.
Something for the buying team
Where demand actually sits across the tonal range, which gaps your range can't serve, and a scheduled export so planners aren't waiting on anyone to pull it.
An engineer who knows your build
One named person through integration and pilot, in a shared channel with your team. We'd rather over-support the first deployment than untangle it six months later.
Start with your own return data.
Bring twelve months of returns and a sample of your range. The baseline analysis is where every honest conversation about shade accuracy begins.