Retail markdown: controlling margin | Key takeaways from our virtual roundtable

Pre-markdown: diving into the data

A common pain point we hear from our merchandising customers is a lack of time (and resource) to take into account all of the data needed to make informed trading decisions. This resonated with our roundtable attendees, with many claiming that data scattered across multiple systems leads to many hours spent number crunching.

One consumer goods key account manager claimed that “we have the data but we aren’t using it. Tracking the decisions and the results in the data is still a massive challenge”

“All of your time is spent on cleansing the data instead of thinking about what you should do with markdown. You can pick up on the slow sellers, but it’s difficult to be proactive,” said another.

A fashion start-up explained that they were favoring “clear as you go” promotional markdown strategies rather than four-times-a-year, big-hitting events. A retail consultant had observed that retailers had become reliant on these quick, fast promotional events – and that “almost nobody defines the objective of the promotion before they plan it. So it’s hard to tell whether it was a success or a failure.”

Markdown strategy

One particularly tough challenge highlighted was the initial planning behind a successful markdown campaign. One attendee cited difficulties around taking into account factors like shelf life, product lifecycles and the need to balance these with “respecting the brand along the way,” plus knock-on effects from the pandemic that continue to linger – such as drastically-altered customer buying behaviors and habits.

Lead times have increased significantly because of the ongoing supply chain crisis, and brands now need to find a way to “incorporate this nuance into their strategy.” Intake margins have risen substantially due to the increase of freight costs and soaring energy prices, which means that retailers now have a lot less margin to play with when marking down.

Peak continually finds that using AI-powered demand forecasting and price elasticities to optimize markdowns are often beyond the imagination of most retailers who are at earlier stages in their data journeys. A multi-channel manufacturer of consumer electricals explained that “we can look at current performance of lines using sell through and rate of sale metrics to highlight slow-sellers but it’s the decisions after that that are missing; it’s the planning ahead.”

One consultant explained that a lot of retailers were still favoring traditional blanket discounting “doing 20%, then 30%, then 70% off. The problem with this is you end up relying on that final hit of 70% – and then it’s too late and you’re selling for less than cost.” Another honestly explained their approach to markdown was more based on gut-feel and intuition; “there’s not much science in it, we just mark it down by 10% and see how it goes.”

Competitor analysis was also a popular discussion point amongst our attendees, with different brands ranking it at different levels of importance when it comes to their markdown strategy planning. One guest claimed that while they looked at what their competitors were doing markdown-wise, no decisions were taken purely based on competitor activity. 

Another, though, believed that competitor analysis was a key piece of the markdown puzzle. “Get your competitor analysis right in the planning stage so your initial pricing is right and the stock right for your competitive position – these are the big wins, everything else is tactical response.”

AI-driven demand forecasting

Another key talking point during the roundtable – and one that is obviously close to our hearts at Peak! – was the utilization of AI for demand forecasting. By leveraging all of your data, not just the top and bottom 20% of a spreadsheet, AI can empower teams with accurate, granular forecasts that allow you to see into the future and react to problems before they happen. 

We discussed the importance of…

1. 1. Data sources: Using AI to go beyond just looking at sales and stock levels, and start to bring in further datasets that give demand explainability. These include promotional/markdown calendars, events and even weather data. 
   2. Techniques and explainability: AI can quickly learn historical patterns and use the data that best explains the demand curves. It also brings granularity at scale, taking into account thousands of product-location combinations. 
   3. Automation and feedback loops: Data feeds AI applications automatically, so there’s no need to pull together data from different systems and spend hours wrangling the data in Excel. AI models constantly learn in light of new data, delivering continually-improving forecasts. 
   4. Customizing AI to your world: An off-the-shelf, one size fits all-approach to AI demand forecasting isn’t going to give you the best outcomes. The forecasting technique, guardrails and ways of consuming the forecast have to work for your business.

Price elasticity

Price elasticity is a measure of how much demand is affected when the price of a product changes. If demand is affected, the product is elastic. Understanding price elasticity allows you to simulate the effects of markdown percentages and choose the point that maximizes margin and sell-through.

Markdown success: the metrics that matter

One major talking point at the roundtable was the best way of measuring the success of a markdown. What metrics do teams consider to be the most important? What results should teams be prioritizing?

One retailer explained that they realized their markdown efforts were affecting gross margin, and that they needed to change their reporting to ensure they could monitor that before and after markdown periods.

When helping retailers with their markdown strategy at Peak, we find that there are four focus benefits of using AI:

• Profitability: Don’t needlessly erode margins by heavily marking down. Make promotional and markdown periods profitable every time

• Reduce terminal stock: Sell-through stock in the selling window and free up capital tied up in aged stock

• Never disappoint a customer: Markdown prices and promotions that entice in customers and keep them coming back for more without ruining your brand perception

• Improve productivity: Empower your teams to spend less time pulling together data and wrangling excel sheets, and more time making informed decisions

Sustainability

Sustainability continues to be a hot topic amongst all retailers, as well as businesses across other sectors. One question raised was around the relationship between markdowns and returns from a sustainability perspective. An attendee had seen that, if you markdown products close to full price, many people may end up buying the product again for the cheaper price – and returning the full-price item. To avoid this, should the markdown, perhaps, only be actioned once a 30-day return period has lapsed? 

This was a big talking point, with one guest claiming their brand actually analyzed why certain items are returned, but that the majority of others didn’t go into this level of detail. Peak’s Retail Director, Tom Summerfield, was of the opinion that markdown can actually help brands dispose of stock in a timely manner, rather than using potentially more carbon to move it around to store elsewhere – with AI also enabling teams to optimize markdown in the best way possible to minimize returns.

Another viewpoint was that in-store markdowns are more likely to be kept rather than returned, as people still love the thrill of visiting a store and finding a bargain!

Conclusion

Our markdown virtual roundtable was a great opportunity for Peak to hear first-hand, from a wide range of retailers, some of the problems they face around markdown strategy – and for them to learn more about the applications of AI for markdown and how technology can alleviate some of these concerns.

Thanks to everyone who joined us to chat all things markdown – and keep your eyes peeled for the next one! If you’re keen to learn more about Peak’s AI-driven markdown application, you can see it in action in our short demo below.