When the user navigates to the Assortment Optimiser screen (by selecting View Assortment Optimisation in the sidebar) they will be presented with a screen showing all the baselined clustering projects. (see image below)
The purpose of this screen is to select a project and start the assortment algorithms to generate an assortment. The assortment dashboard can then be accessed via this screen, as soon as the assortment algorithms have finished their calculations.
When the algorithms are running, a refresh 
button appears next to the project name. Depending on the size of the project, it could take approx. 15 minutes to run, so it is wise to run the project and switch to another task while processing takes place. A (refresh) button will show while the project is processing. Clicking the refresh button allows the user to see the current progress of the project that is being run. (the button will disappear as soon as the project is run)
A downloadable copy of the visual CDT used for each project is accessible by clicking on the project name, and clicking on the download button in the window that appears.
Fig 1. View Assortment Optimisation Screen
- Delete Assortment – Allows the user the delete a project that is no longer required.
- Direct Run Assortment – Select a project (4) and Direct run assortment optimisation. This will bring up a window that allows the user to select a number of variables including the time period to use when running the assortment optimiser algorithm.
- Search Filters – These are filters used to isolate particular projects within the project list. The “Clear” button on the right removes any applied filter criteria.
- Select Project – The arrow can be clicked to reveal a selection box. The selection box is used to select a project, prior to using the “Direct Run Assortment” button.
- View Dashboard – This button only appears once assortment optimisation has been run for a project. This takes the user to the assortment dashboards.
Selecting a project and clicking on Direct Run Assortment will lead to the following dialogue box opening :
- Project Name – Restates the project the user has selected
- Created Date – Date project is created by user
- Model – Algorithm model selected to be used for AO (Normally item recommendation model)
- Discontinued Date – Removes any item from the assortment with a discontinued date greater than the selection.
- Start Date – User defined start of date range for the AO run*
- End Date – User defined end of date range for the AO run*
- No. of Period – Automatically populated after date range has been selected showing how many weeks the project contains
- Planogram Type – Either Standard or Brand. Allows the user to run a branded AO plan if required. Usually Standard will be the appropriate selection.
- Assortment Method – ‘Number of SKU’ indicated the assortment will have space breaks with a specified number of items for each one
- Assortment Type – Past or Future. Past utilises the retailers sales history. Future would be selected if Retailigence was taking a feed of the retailers forward forecasts to enable future AO modelling.
- Space Breaks – The space breaks are shown with the calculated item numbers predicted to fit for each (assumes Retailigence Space Modeller has been run). The user can amend these item numbers.
- CPI (In Percentage) – Allow the user to set the mix of CPI between Value, Volume & Profit that they would like the AO to be run for
*When specifying the start or end date, a warning will appear if the user choses a date that is not a Sunday or Monday. This does not prevent any day of the week being selected by the user.
Notes on Setting the CPI
What is a Combined Performance Index (CPI) ?
- A Combined performance index is a way of ranking a product across several numerical variables that are not easily comparable, such as sales value, volume of units sold or margin.
- An example would be premium headphones, which would have low volume sales but a high sales value and a high margin value. Conversely, cheap in-ear headphones would have a low sales value and a low margin, but high volume of units sold.
- An assortment of products generally has a range of products that vary in sales, units and margin, which makes choosing the best products difficult.
- By using a CPI the variables can be blended into one figure that indicates an individual products performance across the 3 variables (or just two of them), within the selected assortment.
CPI Selection – Blending and Weighting
- With the 3 variables indexed, they can now be combined into the CPI.
- When creating the CPI it is useful to weight the contribution of each variable based on objective/category balance.
- The importance of each variable will change the ranking of the products so it may be useful to test different CPI weightings.
- The weighting allows the user to give a percentage weight to each variable, which should add up to 100.
- The index value for each variable for each product is divided 100 and then multiplied by the relevant weighting percentage.
- By summing the resulting values for each variable for each product, a Combined performance index is created.
- Using the CPI, products can then be ranked based on 3 weighted variables.
Using the CPI
- In categories where there is a diverse range of products and price points using, Sales 50% , Units 50% , Margin 0% should give a balanced product ranking.
- In categories where either volume or value sales is more strategically important then dialling these variables up to 75% may be more appropriate ie.
- Volume orientated category – Sales 25% , Units 75% , Margin 0%
- Value orientated category – Sales 75% , Units 25% , Margin 0%
- Margin should only be introduced into the CPI if the retailer’s profit data is robust.