In my last blog post, I talked about how retailers use traffic data with their workforce management (WFM) systems. The reality is that a minority of retailers use traffic to drive labor.
Why aren’t more retailers aligning labor to traffic? I see five hurdles to adoption.
1. No traffic counters
The largest obstacle is a retailer’s inability to count customer traffic. You need to have some form of traffic counters (also called “customer counters”, “footfall counters” or “people counters”) to monitor traffic in your stores.
Traffic counters are hardware devices that are installed in stores. These devices use technologies such as infrared, thermal imaging, and video to count incoming and outgoing foot traffic. Sophisticated traffic counting systems also offer integrated software that configures the devices, and collects and analyzes the data collected.
The biggest inhibitor to traffic counter adoption is cost. A traffic counter typically costs a couple of hundred dollars on the low end to more than a thousand dollars per store. This doesn’t include maintenance or support services. Nor does it include the cost of the retailer’s internal staff responsible for maintaining the hardware and supporting the data that gets collected.
You can justify the investment in traffic counter by running a small pilot. Integration to your workforce management should be simple. After all, this is just a pilot. Sophisticated, rock-solid integration can be implemented as part of a larger rollout.
After a few weeks, analyze the results. Compare the pilot stores to similar stores that didn’t have traffic counters installed or to the same stores last year. The data will tell you what impact traffic counters had on your store and how quickly additional investment provides pays for the system.
2. Concerns with accuracy
The second hurdle that keeps retailers from using traffic to drive labor concerns the accuracy of the data or, more specifically, the way the counters interpret customer traffic. For example:
If two people walk into the store at the same time, are they individual shoppers or are they part of the same party?
If displays are setup near the traffic counter, does the device interpret the customer examining the merchandise as one customer or will it count him or her as multiple ins and outs as he or she floats around the display?
When associates arrive for their shifts, will the traffic counter view them as associates or customers?
Most gross accuracy issues – including those that I’ve listed above – can be worked through with proper configuration of the traffic counting devices and proper interpretation of the data. The market leaders will guarantee 95 percent accuracy, which is pretty impressive.
Nevertheless, many retailer compare the accuracy of the traffic counters with that of their point of sale system. They decide that point of sale (POS) provides more accurate data and choose to focus on that data to drive labor.
The problem is that this is an apples-to-oranges comparison. Traffic and POS data measure different things. POS data is often used as a proxy for traffic but it isn’t a replacement. If you want to truly align labor to customer demand, then you need to overcome the concerns with relatively small error rates, and use the traffic data that is available to you.
The best way to do this is through some basic analysis to understand how much of an impact your accuracy problems will have on the labor forecast. It has been my experience that small error has a negligible impact on a labor forecast. If you’re seeing significant accuracy problems with your traffic counters, that’s another matter and I suggest you talk to your vendor.
3. Incomplete traffic coverage
Closely related to the previous two hurdles, is the concept of coverage. More specifically, how much of your stores’ traffic activity is captured by the traffic counters? What sort of information do you need about how your customers move around your stores? And more importantly, how does that information affect your planning and scheduling of labor?
Often, retailers have one traffic counter in the store and its located near the front door to track customers coming in and going out. While this provides a good indication of overall traffic, the data provides no indication of what the customer does after he or she enters the store. The customer has, in effect, entered a black hole.
This doesn’t present a problem for stores that have only one department or stores that do not assign associates to a specific department. In these cases, that one traffic counter provides a good indication of how many customers are in the store at any given time and that is enough for labor forecasting purposes.
It’s a different story for stores that have multiple departments and multiple floors in which each department is staff discretely. You have two choices to solve this problem.
First, you can install traffic counters in each department. This presents some challenges from both a cost and a technology perspective, but depending upon the departmental layout and the traffic counters you’re using, it may be possible to accurately count customer flow through each department.
Second, you can proportionally allocate traffic by using sales, transactions or similar metrics to indicate how much traffic should flow to each department. It is unlikely that you can do this manipulation inside of your WFM system. Instead, you’ll probably need to parse the traffic outside of the system and import adjusted results into your WFM system.
4. No traffic-based labor standards
The fourth hurdle to adoption is labor standards. Having traffic data is not enough. If you want to use traffic to drive labor, you need to have labor standards that specify what task is affected by traffic and how much labor is results from each customer counted.
What sort of labor standards are driven by traffic? Primarily, customer service tasks. Customer service tasks include a variety of activities such as answering a question, counseling on size or selection, helping to sample an item, or giving an opinion.
Retailers often blow this hurdle out of proportion because they are often turned off when they hear the words “labor standards”. Labor standards conjure up images of guys with stop watches performing time-motion studies that produce mediocre results that are dated the minute they standards are produced.
If you don’t have traffic-based labor standards, don’t worry. Developing customer service-based standards is straight forward. Most retailers can have such standards in-hand in just a matter of weeks.
5. Not all traffic is created equal
A few years ago, I was working with a North American pharmacy. In their urban stores – those by office buildings – they saw a distinct difference in the needs and buying patterns of customers that shopped during their lunch hour compared with customers that shopped after work.
During lunch hour, customers would come in for a specific thing and their baskets were typically very small. They were buying a soda pop or a bag of chips, dropping off a prescription, buying a greeting card, or picking up aspirin.
After work, customers came in and…well, they shopped. They would leisurely stroll up and down aisles doing a mini grocery shop. If they had a prescription, they would shop while it was being filled unlike lunch time where scripts were either picked up or dropped off. Their baskets were larger and the time spent in the store greater.
Traffic was higher during lunch but sales were higher after work. Customers expectations for service were very different during each period. The labor need was different with the store needing more cashiers during lunch and more associates on the floor providing customer service after work in the evenings.
Similar patterns can be seen in many retailers:
Is the crowd that browses your store before a movie on Saturday night of the same quality as that of the soccer moms that come on a weekday afternoon?
Do your shoppers expect the same level of service during the holidays that you provide outside of the season?
All of this leads to the fifth hurdle: Not all traffic is created equal. Even when retailers overcome the other hurdles, they are often hesitant to use traffic to drive labor because a labor dollar invested during a non-productive period viewed as a poor investment compared with that same labor dollar invested against a productive period. And on their own, traffic-driven labor forecasts will do just that.
Sales represent actual productivity and seem like a safer bet. But as I discussed in my previous post, sales assume certain past results rather than focus on the opportunity which traffic offers.
For retailers that experience these unequal traffic patterns, I recommend considering a blend of labor drivers. You can use traffic as a general indicator of anticipated service volume required and then use more specific drivers to determine where you need to deploy the staff.
What hurdles keep you from using traffic?
The above are the hurdles are the ones that I see retailers struggle with most often when trying to adopt traffic as a labor driver. What are your experiences? Please leave a comment below or feel free to e-mail me if you’d prefer to remain anonymous.
There is a lot more to write about with regards to traffic and its affects on workforce management. Most importantly, I want to explain what retailers can expect when implementing traffic as a labor driver. Look for that post in the near future.