PLATFORM · DRIVER LIBRARY

Every signal behind a price, named and weighted.

A price is never one number from a black box. Elastly breaks it into named drivers, each with a learned weight, a confidence score, and a plain-language hypothesis the model is testing.

Start free trialSee the algorithm
Driver inspectorKVI STATUS
HYPOTHESIS
Key-value items are watched closely, and a mistake here costs more trust than margin.
WEIGHT0.88
CONFIDENCE
High
MAPPED TO
412 SKUs
Counterfactual: drop this weight to 0.40 and the recommended price moves −$2.10.
THE LIBRARY

Fourteen drivers out of the box.

HIGHMEDIUM
KVI status
Key-value items are watched closely, so a mistake there costs trust, not just margin.
HIGH
WEIGHT0.88
Brand value
Strong brands hold their price while commodity lines do not.
HIGH
WEIGHT0.81
Competitor intensity
The more tracked rivals on an item, the sharper its elasticity.
HIGH
WEIGHT0.77
Customer segment
Dealer, retail, and online buyers have different willingness to pay.
HIGH
WEIGHT0.74
Buyer frequency
Repeat buyers notice price moves faster than one-time shoppers.
HIGH
WEIGHT0.72
Price-change acceptance
How a segment has reacted to past changes predicts the next one.
MEDIUM
WEIGHT0.71
Price level
Cheap items absorb increases; high-ticket ones resist them.
HIGH
WEIGHT0.69
Seasonality
Demand and price tolerance swing with the calendar.
HIGH
WEIGHT0.66
Product lifecycle stage
New and end-of-life products respond to price very differently.
MEDIUM
WEIGHT0.64
Product type / attributes
Category and attributes cluster items with similar response.
MEDIUM
WEIGHT0.62
Customer loyalty
Loyal segments accept more; price-shoppers churn.
MEDIUM
WEIGHT0.60
Price-change frequency
Often-repriced items build tolerance; rare changes spark attention.
MEDIUM
WEIGHT0.58
Inventory run-rate
Aged or overstocked items tolerate markdowns; scarce ones command more.
MEDIUM
WEIGHT0.55
Basket size
Items bought in big baskets face less individual price scrutiny.
MEDIUM
WEIGHT0.47
EXTENSIBLE BY DESIGN

Add a driver as data, not code.

Register a new signal, map it to product or customer clusters, and the model learns its weight from your outcomes. No deploy, no data-science ticket.

01
Register

Define the signal and the data it needs, whether a column, a feed, or a derived field.

driver.create("repair_rate")
02
Map

Point it at the product or customer clusters where it should apply.

map → { category: "Tools" }
03
Learn

The model fits a weight from outcomes and scores it for confidence, automatically.

weight 0.61 · conf High
DATA QUALITY

A trust score behind
every price.

Every recommendation carries a confidence score built from the data behind its drivers, and it tells you exactly what would raise it. You always know how much to trust the number.

Sparse-data SKUs flagged, never silently guessed
A clear next action to lift each score
Autonomy only unlocks where confidence is earned
RECOMMENDATION · GaN 65W CHARGER94 TRUST
Sales history depth24 mo
Elasticity fittight
Competitor coveragepartial
Cost-data freshnessdaily
To reach 98: connect a competitor feed for this category, the only driver still on partial data.
DRIVER LIBRARY

See what's really moving your prices.

Connect a data sample and Elastly maps your catalog to its drivers within days, with every weight and confidence score visible from day one.

Start free trialTalk to sales