Labor Costs Are Growing Yet Retailers Haven’t Done This to Fix it


It feels like every other pundit story we read at the tail end of 2016 was noting how hard it was to run a brick and mortar business. Their biggest gripe was labor. These were the recurring trends if we may summarize:

  • Welfare pays better than hourly rates in retail
  • Gig economy empowers workers to earn higher pay than that in retail
  • Once you have employees, keeping good ones is tough
  • Government increases to minimum wage and healthcare mandates make full employment impossible
  • What’s up with robots for labor ails?

Given that we see data from tens of thousands of retailers, we have some thoughts.

First of all, you need to focus on what you can control. Welfare and minimum wage issues are political. If you want to get involved there are ways to do that. But those take a lot of time and you have little guarantee of change.

Second, robots are mostly marketing gimmicks at this juncture. Yes, you can use tablets and kiosks to help customers make orders or find more information, but if your business is trying to be differentiated it needs to do something a quick visit to Amazon wouldn’t (more on that below).

This leaves us to focus on the things that are within our powers.

We’d like to ask ourselves to think differently here. Clayton Christensen proposed something called Jobs Theory and it’s coincidentally not too different from Simon Sinek’s advice to, “Start with why.”

The theory goes that people make hiring decisions for specific reasons. You, as a business owner, might hire people to fill your needs for sales, but the employee is hiring you – as their employer – for their needs. Maybe they need income, want to be challenged, need a sense of belonging, etc.

Taking it one step further, the customer is hiring your product or service for a reason of their own. Do they want your product because it makes them feel good? That they need it for their everyday lives? Is it something aspirational?

Why do people choose to ‘hire’ what they do?

With that in mind, let’s address the controllable portions of your labor.

Labor becomes a huge cost if it’s not appropriately allocated to meet customer demand. Too much labor and you hemorrhage profit. Too little and customers get poor customer service which drops revenues over time… and it’s not as if you’re the only retailer in town.

Using better labor models enables businesses to properly align staffing to meet customer demand. Good models will take into account various externalities – like weather and events – to more accurately predict customer traffic.

Using machine learning for labor optimization is great, and it’s benefits are well documented for retailers. We consistently see 20% profit boots and 1-3% revenue gains when retailers implement machine learning softwares, even when retailers are already using back office solutions that claim to employ “data science”.

But to extract the next level of value, you need to get human.

When customers visit your brick and mortar store, what is their Job Need? If you truly believe that you’re solely a place for consumers to get what they need quickly and leave, you’re going to be out of business soon enough. Ecommerce (and autonomous delivery that brings down costs for to-go orders and returns) will make you obsolete.

If you’re a retailer thinking this way, here are the the leading categories where you’ll remain competitive with ecommerce (at least in the short run):

  • You have items people need ASAP (my toilet is overflowing and I need a part NOW!)
  • You have items people need to touch/feel (what’s this fabric look like in the sun?)
  • You have items where shipping is cost prohibitive (I need 1,000 cheap styrofoam plates that’s a cubic meter in volume)

What brick and mortar retailers should have that ecommerce and delivery services have no answer for is SERVICE. You can create a personal, human experience that has an impression on your customers.

That is the differentiation you can offer. That’s the customer Job Need that you need to solve for. 

And since you can’t be everywhere at once, how do you offer a personal touch?

Your employees.

Your employees are an extension of your culture and brand. It doesn’t matter if you make the best food in the world: if customers are interacting with a total turd for a server they’re not coming back.

Just like customers have a Jobs Need, so do your employees. Your employees want to be appreciated, rewarded and told they’re doing well. Being a top performer but also being consistently overlooked will have any good employee looking towards the door.

That’s why – if we agree that employees are your biggest asset to solving customer Job Needs in the offline world – you need to take care of them. And no surprise, data can do this objectively.

With machine learning rules engines (really a neural network), you can specifically pinpoint the behavioral patterns of employees. You can learn which employees perform well under pressure, and which are losing you money no matter when they work. You can examine training curves and measure impacts of sales contests.

Most importantly, you can objectively reward the good employees by giving them the shifts where they’ll make more money and quickly cut the hangers-on. This boosts employee retention and your revenues simultaneously.

It’s actually not a foreign concept, or at least it shouldn’t be:

ObjectiveLogistics (RIP), a gamification company focused on improving employee check averages, saw a 1.8% increase in business revenues and 11% increase in tips using machine learning to optimize employee performance.

In summary, using data (surprise!) can solve a lot of problems. As it relates to the topic of labor, by using data intelligently your good employees are now making more money (part of their Job Needs), you’re making more money (part of your Job Needs) and customers are getting better, more personal service (part of their Job Needs).

Everyone wins!

Just be wary of reporting tools masquerading as “analytics” that can’t provide accurate assessments of employee performance nor robust labor forecasts. If you find yourself having to interpret too much to take concrete actions it’s not a useful tool, it’s a curse.

Where can you find a software solution employing machine learning to both optimize your labor models and solve Job Needs for your business, all in a prescriptive output that can’t be screwed up? Where else.