Although ecommerce remains strong, the retail sector as a whole faces challenges. Many online retailers that reaped the benefits of the 44% spike in internet shopping during the pandemic found they were left with an overstock as bricks and mortar stores began to open their doors. For most, this has meant relying heavily on discounts as part of their post-pandemic strategy.
However, given how frequent today price promotions are, it is surprising to see that retail businesses actually know very little about their effectiveness. Based on the cross-market data on promotions from retail businesses of all sizes and all types of goods that we at Inventoro.com have analysed, we know that a staggering 67% of discounts result in the business losing money, and only around one in three discounts reflect effective business tactics.
One of the main reasons for these results is that business owners struggle to calculate the profitability of their discount campaigns and account for cannibalisation. Discounts typically boost sales of one discounted item, but this negatively affects sales of other similar items.
For example, a business that puts cheddar cheese on promotion by offering two for the price of one will sell more, but – because there’s only so much cheese that any of us can consume – the sales of gouda, blue cheese, camembert, and others will drop accordingly.
This example is pretty straightforward, but the dynamics of cannibalisation can be extremely complex and unpredictable. Products that may not seem to correlate, such as mountain bikes and skateboards, can affect one another.
Discounted items also cannibalise themselves with pre- and post-cannibalisation effects. That is, customers typically buy nothing in the lead-up to Black Friday, make their purchases while things are on sale, and then go back to spending nothing. In this sense, discounts are just borrowing money from the future.
But the complexity does not stop there. Many argue that discounts may lose money but that this is offset by attracting customer to the store who end up buying non-discounted items. However, the data suggests that this is only true of some items.
While certain products do attract a bigger basket, others don’t – and there is no golden rule to understand which one is which. This “magnetic” effect varies by season, location, brand, and countless other variables.
To fully understand profitability of promotions, retailers need to be analysing the behavior of their whole portfolio as one big ecosystem. This can be only done with smart algorithms, including AI and neural networks.
This effort needs to be carried out at the product level: some products only sell when discounted, some sell better when discounted, and some top-sellers sell regardless of discount. Using data to identify these different segments is key – otherwise, businesses will be losing money.
At Inventoro, we are primarily an inventory forecasting company. The reason why we dive
into promotions so much is that their effects play a huge role on demand. Since it is our role to calculate future demand for our retail and ecommerce clients, we have to analyze promotions very carefully to forecast customer behavior.
In addition to their negative financial consequences, poorly considered discounts are also bad for inventory. Retailers that don’t have the data to discount in a strategic way also lack the insights to determine how frequently and for how long to run these campaigns. As a result, they often run out of inventory, which is a tragedy for both profit and customer service.
We at Inventoro.com are far from advocating that discounts should stop altogether: price promotions are a core part of capitalist consumer behavior and ending them is unrealistic. Instead, we advocate for data-driven smart promotions. With the use of advanced algorithms, AI and deep learning principles, retailers can calculate the optimal balance between discounts, profits and inventory.
Radim Jung, co-founder, Inventoro.com