Skip to content
Contact us
  • There are no suggestions because the search field is empty.

Non-recurring Events

What is considered as a “non-recurring Event”?

The list of cases that should be tagged through the Events module include:

  • Events that don’t happen every year (non-recurring)
  • Changes in demand that impact a small scope of products within a family (weak signal for the model to capture automatically)
  • Changes in demand that impact a given store or a small scope of stores (weak signal for the model to capture automatically)
  • Changes in demand you are unsure the model is capturing automatically (due to lack of data quality, proper attributes or inconsistencies on historical data)
  • Changes in demand the system would otherwise be unaware of (concerts, sudden discounts, marketing campaigns, festivals, strong weather changes, ad-hoc promotional activities, etc)


How to create non-recurring Events in the Platform


To add a new Event into the system go to Master Data → Events. Here, you will see cards for Ongoing, Future and Finished events that have already been created. On the right hand side, you will see a button that says “New Event”:



Here below some guidelines to fill up all the necessary fields when creating a new Event:

  • Event name
    Provide a clear, identifiable name that the whole team can easily recognize and that will help you differentiate between all the events created in the system. For example:  “50% off WW Shirts Germany March 25”
  • Input the Product and Store Categories impacted
    Select the Product Category that includes the scope of products that will be part of this event (in this case, all the Womenswear Shirts that are getting a 50% discount during March), and the Store Category that contains the stores where the event is actually taking place (in this case all the stores in Germany)

In case there are no categories currently created in the system that contain the specific scope of products/stores of your Event, you can create them beforehand as explained in the next chapter:  “Categories”.

  • Indicate the Period
    Choose the Start Date and End Date of the event. These dates determine the period during which the event is expected to impact the demand trend of the products and stores selected in the previous step.
  • Event Coefficient
    Indicate the expected increase in demand that this Event will have in the products and stores selected. For example, if you anticipate that this discount will increase the sales of shirts by 45%, you should set the coefficient to 1.45.
  • Create the Event
    Once all fields are completed, click on Create.

Extra configuration possibilities:

  • Assign a Similar Event
    In the “Similar event” field, you have the option to use a past event (from previous months or previous years) that is similar to the one you are setting today (same discount strategy, same product type, same stores…). Once selected, the average coefficient of that past Event will come up - and that’s the one that will be used for this new Event you are creating today. 

In case you feel this year’s Event will have a higher/lower impact than the one from that Similar Event, you can simply overwrite the coefficient.

  • Event Type
    In order to give more information to the model about the type of event you are creating, you can do so by selecting any of the following tags:
    • Marketing campaign: if the expected increase in demand comes from marketing activities you can use this tag (banners on the web, changes in the shop floor, window displays, collaboration with influencers, tv advertisement…)
    • Discount: this event consists of direct price reductions (x% off)
    • Works in store: this event type is used to tag periods during which the store might have to be closed for a given time (weeks / months)
    • Circumstancial: this event represents how external circumstances might affect the demand (concerts, city/neighbourhood events, festivals, weather changes, school events…)
    • 2x1: instead of running direct discounts, this event will consist of conditional offers (buy two get one free, spend more than 100€ and get 40% off…)
    • Store resite: when a given store changes its address and code, all the historical data from the previous one can be copied to the new one (ad-hoc request). However, the system is going to be unaware of the potential increase in demand that might be generated right after the reopening due to that “novelty effect” and footfall increase. 
    • Bet: all those events that are created to increase the demand based on the user’s knowledge/expertise - even if no promotional/tangible action is behind this increase  

The reason all these tags exist is to be able to give the information in a structured way to the forecast model so that it is possible to learn from each of them by establishing the correlation between each event type and the actual impact they tend to have on sales. This has been created with the long-term vision of having the model proactively suggesting a coefficient when the user creates an event (rather than having the user input it). However, for that, we would need to have hundreds of events tagged - so that the density of data available is high enough to start learning stable patterns. 

As a consequence, there is no short-term impact when assigning a given tag to each event created. It would only help for a long-term strategy (hence why it is offered as an optional field only - so that each customer can decide).

 

How is the Event information used by the model once created?

The coefficient will multiply the daily demand of all the product-store combinations that are part of the event created. 

Let’s see it with an example:

  • Product Category: Product A
  • Store Category: Store B
  • Event period: from September 4th to September 10th
  • Event coefficient: 1.70
  • Daily demand (simplification) of product A in store B: 5 units
  • Lead Time of store B: 3 days
  • Planning Horizon of Replen: 10 days
  • Replen Execution date: September 2nd

As the daily demand is 5 units, if no promotion was created:

  • The Stock Forecast (lead time) would be: 15 units
  • The Demand Forecast (PH) would be: 50 units

As the daily demand is 5 units, if this promotion (1,70) was created:

  • The Stock Forecast (lead time) would be: 22 units
  • The Demand Forecast (PH) would be: 67,5 units

As a consequence, it is important to be very careful when setting up events. If one product is included in two events at the same time, its demand will be multiplied by both coefficients (potentially generating a very high demand).

The replenishment screens will give you visibility on how many events a given product has applied during the same period (these screens are explained in more detail in the chapter 5 of this course), but here a screenshot:




Once the event period is over, the system will calculate what the average impact in sales was by comparing the units sold during the event period against the units that were sold during the 3 weeks before.


When should I configure my event in the system?

The short answer is: the sooner the better. The best approach is always to configure your events in Nextail as soon as you receive all the information (product scope, store scope and period). This way it will be considered by the system as soon as it enters within the replenishment period (lead time or planning horizon).

Otherwise, the guideline to assess when the latest an event should be configured is calculated as your Planning Horizon + Lead Time.

Let’s imagine there is an event starting on September 30th that impacts store A, which has a lead time of 6 days. If your current Planning Horizon was of 14 days; the latest that event should be created in Nextail to ensure the store is prepared on time would be on September 10th.

 

How to create a batch of events using an Excel File


In some cases, the user might need to create several events at the same time. For example, in Black Friday different discounts might be applied to different product types and therefore the sales uplift expected will vary between them. As a consequence, rather than creating a single event with an averaged coefficient, creating one event by product type will most likely give better results. Whenever this is the case, Nextail gives the possibility to create a batch of events across different store and product categories through a Data Upload.

 

  • Obtain the Promotions Template and fill it up

If you don’t have it already, remember that you can download all templates from the Knowledge Base. Click on your user name icon in the platform, go to the “Help” section and search for “Data Upload Templates”. Click on “Promotions” and the file will be automatically downloaded.

Now ensure your file contains exactly eight columns - named as shown in the screenshot provided below. Each row in the file needs to represent a unique promotion and it has to be filled following the guidelines explained on section 2 within this chapter:


Note 1: The data upload is case sensitive so ensure you are using the exact same product and store categories as they have been created in Nextail (otherwise the upload won’t go through and the event won’t be created).

Note 2: All fields must be filled up for the data upload to work except for “features” and “similar_promotion”, where the user can specify the name of a past event if they would rather use its coefficient. If you leave this field blank, Nextail will default to using the past 60 days of data to calculate the promotion impact.


  • Upload the template
    Go to Daily Data > Data Upload, select the Promotions data type and choose the Excel file you just created to upload:


Following these steps make it possible to create several events  (all at the same time) rather than having to create them one by one on the platform.

Nextail will also see the warehouse stock available for all references of the chain and consider it all as available when replenishing that Link Line to the store network