You’re a retailer, and it’s a Tuesday. You’ve got 17 marketing campaigns running instead of the usual 12. It also happens to be in the middle of summer holidays, but it’s unusually rainy for the season. Taking all of that information into account, what kind of sales and transaction performance should you expect? And are you doing better or worse than that?
What most retailers wouldn’t give for a clear-cut, data-driven answer to those two questions! There tend to be a lot of ‘I reckon’s when it comes to judging the impact of different variables on sales performance, especially for retailers juggling a complex web of brands, stores and campaigns:
‘I reckon the spike in sales had to do with school holidays.’
‘I reckon the drop in store visits was due to the poor weather that day.’
‘Wasn’t there a cruise ship in town - and didn’t the All Blacks play last night? I reckon that’s why sales slumped by 4%!’
These speculations might provide enough peace of mind to help you sleep at night, but when it comes to making promotional and operational decisions, relying on speculation can lead to more questions than answers. Especially when it comes to knowing what action to take, or when to panic.
Are you relying on ‘I reckon’s at your company?
Maybe you aren’t - maybe you do use your data to create some averages around which to base your definition of success. But if you’ve got dozens of stores of different sizes and demographics and you average the impact of a promotion or other variable across them, the result will be wrong (on all counts). Not every day is average, and an average day at one store is not the same as another.
In order to actually understand how different variables (e.g. store size, campaign type, weather, season or location) impact retail performance, you need more than a simple calculation - what you need is a robust and reliable baseline to compare results against. It needs to be comprehensive across many different situations and take meaningful variations into account, otherwise it won’t provide an accurate benchmark for you moving forward.
Having a baseline like this is integral to deeply understanding what’s working and what isn’t, but many retailers don’t even know where to begin because it seems implausibly difficult to achieve.
‘So how can we get an accurate sales baseline across multiple retail stores and metrics?’
Believe it or not, you can understand what expected performance on a specific day at a specific store running specific campaigns should be. Datamine does this with our proprietary ‘Big Wave’ process that uses historical situations and variables to create a baseline that can look at a specific situation – both to determine what should have happened, and how important different factors are to the result. Sometimes we are tracking and modelling thousands of metrics a day across different stores and metrics.
This solution can be used to move the needle on your results in a number of different areas:
- Marketing: Big Wave is particularly useful for large retailers that run many campaigns simultaneously and don’t know what the impact of each promotion actually is - being able to isolate the effects of different campaigns allows them to put money in the best place
- Operations: Big Wave also helps organisations understand (from a logistics point of view) how different variable combinations can affect potential events, such as expected sales, volume of calls or number of expected faults in a network
- Market share: when used in conjunction with the lens of market share information, a Big Wave model is even more powerful – a decrease in sales might look negative at the outset, but if your market share is going up, there is less cause for panic
If averages and ‘I reckon’s aren’t enough anymore, we’d be happy to talk to you about potential solutions in a free phone consultation - schedule one by clicking the banner image below. In the meantime, check out a couple of case studies outlining the Big Wave models Datamine has created for our clients over the years:
One of our entertainment clients had about 20 campaigns running at any one time and they wanted to optimise their marketing spend. However, because they weren’t sure which campaigns were working and which weren’t, they were hesitant to stop running any of them.
In order to give them insight into the issue, Datamine built series of explanatory models that used historical data to isolate the impact of particular campaigns, highlighting a small number of promotions that were seeing the most success.
This information gave the client the confidence to move investment into the things that were actually working, improving their overall marketing ROI.
Another client, a large retailer, had a massive ongoing campaign running but didn’t know how it was impacting different locations throughout the week. They needed to know the impact of the promotion on both a store and day-by-day level across the network.
Datamine created Big Wave models that took the different variables (such as store location, competitor promotions and day of the week) into account, helping the client understand both total actual incremental impact and the likely logistics and distribution changes they’d have to make.
In addition to providing insight into how the promotion performed in different places and on different days, the model showed the client the impact their competitors’ promotions had on their sales. They adjusted the promotions with this information to gain competitive advantage and get more from their marketing spend.