Why does it matter?

Sale periods like Black Friday are challenging for merchandising teams. They can be make or break for many businesses and all the usual rules of the game change.

Products that are usually best sellers can lag behind if similar products are well priced, stock position changes rapidly, and marketing efforts can introduce surprises in traffic and demand for certain products.

Regular targets around first price sales mix, margin etc can go out the window as the focus shifts to clearance and acquiring new customers. 

Monitor: Start by understanding your No Result rate in your Analytics 

Fundamentally there are two main drivers, what customers want to buy and what the business wants to focus on selling. You need good quality data to support both and it needs to be fresh. Looking at a sales report for yesterday in Excel and adjusting your merchandising today is too late, plus everything may have changed in terms of how much stock you have. Further discounts may have been applied since then too, so knowing how a product performed yesterday isn’t as useful as it would usually be. 

Even AI and ML tools can struggle to adjust quickly enough to the sudden change in customer behavior that a big sale causes. A model trained on the last few days or weeks data can rapidly become out of date and it can be challenging to re-train when prices, stock, and behavior are well outside the norm.

You’ll likely have a few leading products, those that have deep stock and a big discount. Manual merchandising works fine for these but it can be tough keeping up with availability especially across multiple regional store fronts and outside of regular working hours.

Having up to date information on stock, discount level, recent sales and any other key business metrics is important all the time but even more so during sale periods. What’s just as important is feeding that information directly into your merchandising algorithm as frequently as you can. 

Algolia customer Gymshark is known for their Black Friday and mid summer sales. AI driven merchandising helps them drive sales all year round but during Black Friday the deep discounts and massive spike in traffic present a unique challenge. The old approach was to have the trading team continually download and merge reports of sales, stock positions etc and adjust the pages accordingly, but for Black Friday 2020 they decided to take advantage of Algolia’s high speed and simple APIs to close the loop and automate a big chunk of the work.

They started by looking at the Excel based process the trading team were using, identifying the data sources, weightings and priorities that went into it. Next they built a prototype workflow in Alteryx to mimic the process and output a business priority score from 1 to 10 for each product. 

That score was then injected in to Algolia via a simple API endpoint directly from Alteryx and added to the ranking formula, allowing it to be used along side other data, rules, personalization tools etc.
By recalculating and re injecting the business priority score every 15 minutes Gymshark were able to automate the majority of their Black Friday merchandising across 14 international store fronts, allowing the trading team to focus on specific areas where their domain knowledge added value. 

Conclusion

Planning for a good merchandising strategy for optimizing business metrics during crucial sales periods by looking at no results rate in your Analytics data.