Retail is a very low margin business. Some retailers scrounge higher margins in commodity goods sold at a premium via branding (think high end fashion, cosmetics, and fragrances) but many are not so fortunate, especially with the entrance of ecommerce a la Amazon.
Restaurants are luckily insulated – to some extent – from the ecommerce competition seen in retail. This silver lining has nothing to do with their business acumen so much as it does good fortune: meals are perishable and must thus be made locally. However, their saving grace is also a major problem in that manufacturing meals requires more operational complexity than listing a non-perishable on a shelf. And since restaurants struggle with even the basics, including giving away their customer relationships, restaurants almost universally earn lower margins than retailers.
So you’d think retail and it’s brother, restaurants, would be eager to optimize their businesses. Of course if that were the case we wouldn’t see such abysmal failure in these industries, especially when some obvious fixes aren’t being executed. This is, in our best estimate, a demographic problem: the majority of people who end up in retail and its concentric verticals don’t end up there after blowing out the doors at a top tier company.
It’s this same “people” problem that keeps a lot of good things from getting done. But what if these “people” were no longer the bottle necks? In this article we’re going to assume rational behavior (perhaps forced by shareholders) prevails and the industry transforms itself to align with market inertia.
We hope it’s no secret that data, and analytics on top of the data, are making big headway in the market. Companies using information more intelligently are getting better returns through shrinking expenses, growing revenue, or both. In the profit-poor industries of retail, these sorts of techniques cannot be ignored.
If you’ve paid attention you’re already seeing the impact of data in brick and mortar competition. Merchants are most likely ignorant as to why they’re using these tools, but if you read this blog frequently it should be clear why the below examples are in place:
Retailers offering omnichannel purchase methods. A digital channel allows the merchant to collect much more data on the user and more quickly A/B test pricing and promotions. This can be fed back into the retailer’s in-store assortment.
Ordering kiosks implemented in restaurants. These cheap pieces of hardware increase check averages and can do a better job of upselling via algorithms than your (soon-going-to-$15) hourly worker. Table-side tablets speed up service and reduce the number of front of house staff needed.
Obviously there’s a lot more than can and will be automated as the industry faces increasing pressures from new channels, regulation, and competition. We’re breaking up these inevitabilities into different categories so they’re easier to understand.
Operations
- Human Resources: not only will data be used to better identify and screen candidates, it will also be used to properly structure incentive programs. What employees are performing the best? Is there data indicating someone should be promoted, demoted, or moved? What about employees who are showing signs they might churn? Expect software to replace a good number of HR employees in the org chart.
- Labor: machines will build schedules anticipating customer demand and incorporating employees’ individual quirks. Ridiculous logic accommodating state and local labor rules will also save time and eliminate compliance concerns. Should certain employees only work certain days? Are teams of people over or underperforming other teams? Expect software to save manager time and also eliminate operations employees at multi-store groups.
- Inventory: inventory will be ordered based upon forecasted ingredient demand while monitoring price fluctuations from suppliers. Automatically. New items can be recommended using aggregate data and recommended pricing can be inferred. Machine vision and sensors will eventually establish par counts at restaurants, saving enormous amounts of time. Say goodbye to most of that supply chain team.
- Hours of Operation: hours of operation can be appropriately tuned to meet foot traffic patterns. This means taking advantage of well-performing hours to boost profitable revenue, while curtailing extended hours that are unprofitable to the business. This task falls to operations folks and the software solution is much more accurate here.
Marketing
- Pricing: sales data showing item relationships, margins, and pricing will allow software to recommend pricing changes for items. If your marketing department is sophisticated enough to do this today, they just found much of their work replaced by some simple analytical software. More savings for the merchant!
- Promotions: don’t know what promotions to run? Now software can run simulations and use more aggregate data to recommend promotions that have a higher chance of working. Cohorts can be built and A/B tested automatically without interference. Iterative maximization happens, and we coincidentally just cut a number of overpaid employees while increasing business revenue. Did we mention how awesome software is?
- Upsells: much like the price elasticity example above, clustering algorithms can help determine what items to recommend to customers. You’ve probably seen these when checking out at Amazon and they say “customers who bought this also bought this.” Kiosks can do this well if management is smart enough to get out of the way, or specific reminders can be sent to front of house employees during checkout or ordering.
Real Estate
- Site Selection: this is a huge component of store success. Would it surprise anyone to learn that data is incredibly important when making this decision? Machines can analyze complex data and find relationships much faster than a human with the best software. Software can make prognostications and build models much better than a person. Your real estate department has been reduced to someone executing what the software outputs. That is, until machines can buy and sell without human intervention.
Now, there remain areas where humans are paramount. Customers visiting a local store likely prefer human interaction otherwise they would just purchase goods and services remotely. Thus the solutions discussed above provide a means to reprioritize employee time to focus on the customer and not mundane “back office” work. There are just some things machines can do better than us so why don’t we let them?
If you think this sounds too expensive let us give you a quick example that proves otherwise.
Aryb’s is a massive brand with over 3,300 locations. Of those, about 1,000 are company-owned.
Using Linkedin, there are about 650 Arby’s employees in the Atlanta area, where Arby’s is headquartered. If we filter for those in marketing, only 20 results show up. This feels really low, so our guess is that Arby’s has creative titles that obfuscate the roles for a number of these people.
Our guess is there are 100 Arby’s employees in the marketing division. Their direct overhead, plus their ancillary budgets to purchase subpar tools from friends in the industry, is likely $25MM a year. Software could very reasonably replace 85% of those employees at much, much lower costs while increasing ROI and proving it quantitatively. The savings can be passed through to customers, and over time make Arby’s more competitive.
There remain numerous opportunities for merchants. Will they execute them anytime soon? Not without shareholder intervention; it’s going to take competition from virtual restaurants and expanding companies like Amazon before more effort is expended by traditional merchants. But the impact is huge for the income statement.
Unfortunately the status quo equals job security for many of these merchants. And that is still the real barrier to this progress.
Add comment