We are deep in the GenAI hype era, so it’s easy to forget about that other kind of AI – the optimization kind. Remember when personalization was THE red-hot AI? It wasn’t that long ago. And while it has long had impact for eCommerce sites, personalization has always been challenging to deliver into the bricks & mortar context – and it will be even more difficult this holiday season in particular.
What is involved in in-store personalization?
Before diving into why, it’s probably worth level-setting on how in-store personalization might work. On eCom sites, you don’t need to know anything about a shopper in order to use personalization to at least some effect. As a shopper browses a site, things like their search terms, their navigation choices, even the filters they choose to apply can be used to better understand their intent and to help support that intent with personalized recommendations.
Those can be applied on category pages – what goes into the default order of items that is presented to the shopper – or it can be applied on product detail pages, with recommendations for complementary items or other similar items in the category that shoppers might choose if they don’t like the item they’re currently viewing.
None of these use cases are available in stores. You can potentially take a specific customer’s online shopping history and use it as a basis for recommendations in a store, but then you’d have to know who the shopper is, in order to apply it. You can’t use the customer’s in-store browsing history – it’s just not something that the vast majority of retailers can collect, let alone act on it in real time.
You can make complementary or substitution recommendations in a store – a well-merchandised store makes those recommendations through which products are displayed together, it’s kind of a base expectation of what stores do. But to make personalized recommendations is nearly impossible without identifying the customer first, and there are a few more concerns that retailers must address before even attempting in-store recommendations.
So the ideal use case for in-store personalization would involve a customer coming into a store, engaging with an employee who has a mobile device and volunteering their personal information so that the employee can pull up that customer’s information. The employee would then use that information and the real-time conversation with the customer to drive some personalized recommendations.
In-Store Personalization is Hard – Really Hard
That in-store scenario seems relatively straight-forward, but almost every single one of those steps holds big challenges:
How will you prioritize local inventory?
Do you recommend items that are a high match from a personalization perspective but out of stock in the store, or do you recommend items that are available in the store but are not a high recommendation match? If you score it more towards personalization – at least you can capture a sale, even if the inventory is not in that store – you’re running the risk of store associates not using the recommendations because they want credit for sales of inventory in that store. Plus, you just spent a lot of time and effort getting that customer into the store, wouldn’t you want to try to sell what’s there first?
On the flip side, if you recommend things that aren’t a good match, then over time the customer may start to feel like you’re not relevant or that the brand “doesn’t know me.” There’s no right answer here, and no real data out there as to which approach is more effective – and it may come down to balancing short-term gain against long-term retention, something that you might want to move back and forth over time to prevent over-indexing too far in one direction or the other. Of course, if you know which of your shoppers are most likely to shop at a specific store, then you could possibly create an assortment that better caters to their categories, sizes or color preferences out of the box – but that is a totally different kind of capability than personalized recommendations.
How are you going to work the store associate into the delivery of recommendations?
There are a few parts to this challenge. First, it’s a loaded question because it presumes that the store associate must be a part of delivering personalized recommendations. It’s true that you could bypass the store associate. You could have in-store recommendations available to the consumer on their mobile device, for example. It would require reminding the shopper that they could get those recommendations while in the store, and might require having the shopper identify which store they’re in so that you could pull local inventory.
Why would you want to do that? Well, having a store associate deliver personalized recommendations requires two things: knowing enough about the shopper to access personalized recommendations (as opposed to generic ones), and having the permission or opportunity to deliver said recommendations in a way that is invited by or accepted by the shopper. I don’t personally crave an experience where I walk into my grocery store, get greeted at the door by a person who tells me “Welcome Nikki! It’s been almost two months since you bought toilet paper, did you add it to the list?” That kind of personalized recommendation is creepy and unwelcome.
But again, why would you want to do all that work to drive a customer to a store – where you pay for the space, the inventory, and the people to be there to create a welcoming environment and engaging brand promise, only to have to put up signage in the store driving the customer to their mobile phone? Every time a customer pulls out their phone in a store is an opening for a competitor to steal your customer away.
Privacy concerns or not, the store associate is the best way to deliver personalized recommendations in stores. They know what’s in stock and what’s not. They know which products are problematic for fit or quality, and they can best guide the customer in the moment – plus they have an opportunity to not just guess the shopper’s intent, like online personalization does, but to actually ask the shopper directly about what they’re trying to accomplish. Even better, they can then use that knowledge, the personalized recommendations, and their own experience to create new recommendations on the fly.
But you need those two essential elements to make that happen: you need to create a situation where the customer opts in to identifying themselves, and you need to have the training and the technical support to ensure that the store associate knows how to incorporate those recommendations into the conversation in a way that feels helpful and delightful to the consumer, and not feel exposed or creeped-out.
What role will mobile technology play in delivering in-store recommendations?
This is another loaded question, with the key word being “mobile.” There are multiple mobile form factors that might be used, but to be successful with any kind of recommendations, they need to be delivered before the customer gets to the till (once they’re at the register it is simply too late to impact that trip’s decisions). And if they need to be delivered by a store associate, there’s just no way for them to access anything (especially if you want it to be up to date in near real time) without a mobile device.
Fine, easy. Get them all mobile devices! Well, if only it were that easy. Most retailers still have not perfected the right device for the right situation. The ergonomics are challenging – a store associate only has two hands and if you require them to always have a mobile device in one, then it limits other things they can do. And if they set a device down somewhere, you need to make sure it’s either not a very desirable consumer-grade device (in which case, store associates may not want to be seen using them anyway), or you need to make sure you’re protected if the device accidentally finds itself in a customer’s pocket on the way out the door. If you give them a sling bag or fanny pack, they might not want to wear that either, or they may just as easily slam that bag into a rack or a doorway, damaging the device. The physical realities of managing mobile devices are still very challenging.
The other side of mobile devices is the data side. How much customer data is okay to share with store associates? How much data is okay to load locally – which makes it more reliable to access – on a device that could potentially walk out the door? And how connected is your customer data to your point of sale? Can you easily go from recommendation to a sale, without requiring the store associate to start over from scratch in building a basket? There are no easy answers here either, and lots of still-emerging practices, especially around store associate access and use of individual customer profile data. But you can’t have personalized recommendations without “personal” – so one way or another, retailers are going to have to figure this out.
The holiday season makes all this worse
Take average traffic and sales volumes and multiply by three, and that’s probably a fair guess as to the difference between holiday vs. any other time of year. But this holiday season will be especially problematic, thanks to the continuing labor crunch. Either retailers will just staff down a level in order to balance out rising wages that have bumped up against a labor dollar budget, or they are going to hire a bunch of seasonal workers to staff stores – workers who are not going to have the time or training to figure out how to elegantly work identifying a customer or making personalized recommendations into a sales process.
Just in general, holiday is a problematic time for anything personalized in stores. There are fewer triggers or events that help incent shoppers to identify themselves before they get to the till. And there are fewer store associates available per customer, just because the holiday traffic is generally much higher than other times of the year, so the opportunity for lengthy, personal interactions become much smaller.
About the only opportunity a retailer may have to get personalized during holiday would be if they were willing to pull a recommended item to go alongside an omnichannel store pick-up order, as a suggested upsell. But this assumes a couple of things. One, that you have the space to store more than just the items already bought and paid for. Two, that you have enough inventory of recommended items that you can risk pulling an item from the sales floor as a “maybe” upsell and not risk taking it out of availability to a customer in the store who may be looking for it. Three, that you can make that upsell opportunity painless and not a whole other order with all the ringing up and payment requirements that a store pick-up order is specifically designed to avoid. And four, that you have a process for handling all the refused upsell items and getting them back on the floor quickly. If you don’t have a streamlined process for this kind of thing already, then it’s too late for Holiday 2023.
The Bottom Line on In-Store Personalization
Personalization will come to stores. The challenges that exist today are difficult, but not insurmountable. And retailers’ wage and labor pressures aren’t going away any time soon, so they face a lot of financial pressure to invest in things like mobile technology – and personalization is one use-case that could potentially help pay for that investment.
But in order for personalization to be successful and not creepy, you need two key things: a compelling reason for a customer to opt in to being identified long before they get to the till, and training and technology to help the store associate make the most of how to deliver personalized recommendations.
But most important, retailers are going to have to decide for themselves whether they are going to prioritize recommendations that have a high score for that customer, even if the inventory is not available in the store. Or if they are going to prioritize in-store availability. The ideal situation is to have both – a high score and available. But stores just don’t have the same level of inventory availability as eCom, and yet they have a much higher expectation for delivering instant gratification. Just make sure that whatever you decide, you do so knowing that you are having to balance between the short term of making the sale against the long term of customer delight and loyalty.
It’s still a long road.
Read the full article here