Current Engagements
Just a few of the organizations we have worked with to grow business through better decisions.
ALLOCATION
(Specifically Size Curve Optimization)
Results: Able to accurately predict over $500M incremental growth by model driven store size / color / pack allocations.
Projected Incremental Impact: Over $500M
DEMAND FORECASTING
(Competitive Test Versus SAP/Oracle/Manhattan)
Results: Beat closest forecast competitor by 66% forecast accuracy after 11 days of calibration, all bother vendors had over 30 days to calibrate models.
Projected Incremental Impact: Over $500M
CLUSTER • ASSORTMENT • PRICE
(Specifically Size Curve Optimization)
Results: Leveraging clustering, allocation, assortment planning, assortment optimization, pricing curve, pricing optimization models – both independent and constitutive forecasting methods.
Projected Incremental Impact: Over $500M
MARKET TREND PREDICTION
Results: 2020 – 2021 markets predicted, consumer trend insights captured. Path to captaincy where there was previously none.
Projected Incremental Impact: Over $100M
COMPETITIVE TEST ON DEMAND FORECASTING MODELS
Results: Testing forecast accuracy over the next 9 months. Currently recognized as most advanced forecasting technology as evaluated by partnered Educational institution.
Projected Incremental Impact: Over $500M
MACRO OPTIMIZATION
Results: In testing to provide recommendations on store format, department and category size optimizations, and adjacencies.
Projected Incremental Impact: Over $300M
CPG Use Case
Opportunity identification.
WHAT WE DID
Company X wanted a viewpoint on category trends over the next 2 years. Using a combination of company sales and KPI data – along with market, demographic, climate and census data from the last 5 years – we ran our engine to determine the best predictive forecast per category.
REFINEMENT
We then back-tested the top models to see how closely the models would have predicted events over the last 5 years. This included testing against major holidays and events to understand how behavior changed.
RESULTS
With 2020 and 2021 markets predicted, we can now understand key behaviors, clustered geographies, and buying patterns by category. Using this forecast, we segmented the buyer types and determined how to influence them based on the trends.
Retailer Use Case
Head-to-head demand forecasting.
WHAT WE DID
CW wanted to predict the impact of planning and forecasting decisions within the vitamin category to see if there was enough financial impact in changing their current approach. Leveraging product sales, allocations, and attribute data – along with demographic, climate and census data – we forecasted the optimal SKU/store/inventory combination.
TESTING
Over a 4-week period, we then went head-to-head with the existing method of forecasting. We were measured based on our weighted daily average distance between our forecast and the actual results.
RESULTS
We were able to prove the accuracy of our engine, beating the existing process of prediction by 58%. Through this, we were able to identify a significant, multi-million dollar impact through changing this process, without even altering the upstream decisions (i.e., assortment, pricing, promo). Insite AI did all of this with only 6 days of AI modeling, and the expected accuracy will continue to improve over time.
