Improving the Operational Art of Curbside Pickup and Other People-Based Processes

Introduction

These are times of extraordinary disruption. Processes developed over years or decades are suddenly being revamped, re-thought or simply discarded. Often to be replaced by significantly more complex processes that must be put in place with limited technological support in what seems like hours. Curbside pickup is a great example of a complex process retailers had to implement almost overnight. Naturally, this hurried march to execute a complex process leads to a lot of bad customer experience. But bad customer experience is something that retail – already struggling with accelerated shifts to digital – simply cannot afford and that no business should tolerate. There are no magic solutions to putting hurriedly designed, complex processes in place at hundreds of physically separate locations with part-time and largely untrained staff. It’s going to be hard.

The Key Problems

People-based processes aren’t fundamentally different than any other kind of process. But there are three aspects to people-based processes in retail that are particularly challenging:

Lack of Visibility: Most people-based processes have no way to measure what the people are actually doing. This makes it nearly impossible to drive process improvement.

Demand Variability: Unlike manufacturing processes that feature relatively continuous flows, most retail people-processes have to adapt to wildly variable flows. Overflow conditions are a common pain point.

Communications: People-based processes in retail usually involve significant customer communications. When combined with high demand, communications systems are particularly prone to breakage.

Curbside Pickup : An Anecdotal Case Study

Curbside pickup is a paradigm case of the kind of dramatic operational transformation necessitated by Covid-19. It’s a complex process involving lots of new Associate tasks, it requires significant customer communications, and it can be a make or break experience for businesses trying to adapt and retain or build customer loyalty. A lot can go wrong. My first real-world experience with a leading home improvement company picking up garden soil and supplies is a perfect example.

I’d placed a curbside pickup order online on Friday and every part of the digital experience was excellent. Store inventories were known. Pickup expectations were appropriately set. I was notified Saturday morning via both email and text that my pickup was ready at the store and I was given a clear order identifier and instructions. Like most digital experiences, it was smooth and efficient with plenty of customer communication.

When I drove into the store parking lot Saturday afternoon, things took a sharp turn for the worse. There was good signage directing me to the curbside pickup spots, but the lot was jammed, and all the spots were taken. There was no visible signage for what to do. I parked as close to the curbside pickup area as I could and walked over. There were signs at each numbered parking space with the curbside pickup number and extension to call. Ok. I called. And was put on hold. After five minutes, an operator came on. “How can I direct your call?” Which, I thought was already done since I had entered the curbside pickup extension, but…”Curbside pickup, please”. Back to hold.

After a few minutes the line picks up but there is no voice. “Hello? I’m here for curbside pickup.” Back to hold. Another few minute wait and it happens again. Then back to hold for another five minutes. The Operator. “How can I direct your call?” At this point I’m a little frustrated but still in polite zone. “I’m trying to get curbside pickup.” Back to hold. Another session with the pickup and say nothing folks. I’m thirty minutes on hold when the operator picks up for the third time. “How can I direct your call?” “I’ve been waiting for thirty minutes to speak to curbside pickup. They aren’t answering,” I say, frustration undoubtably leaking out of my voice. “I’m sorry,” she says, “There’s a tent in front of the spots, go there and give your order number to one of the people there and they’ll take care of it.” Now I have seen this tent. There are a half-dozen customers milling around in it, but there have been no Associates there in the past thirty minutes. So I hang up and leave.

That’s a bad experience in every respect. Terrible operations. Terrible communications. Terrible planning. It’s hardly unique. About half of the curbside pickup experiences I’ve experienced ranged from bad to awful – and almost always when it was obvious the underlying process was overtaxed.

The Solution

Making these experiences better requires solving the three big people-process challenges:

Add Measurement: Deploy technologies that track people during a process and use the data to optimize the underlying process

Create Overflow Strategies: Design and implement strategies for high-demand times when aspects of the process are necessarily overtaxed. People-based processes always feature at least one constrained resource (people), but in the curbside pickup example, parking spaces also lacked an overflow strategy.

Embedded Communications: Customer communications is often the first thing sacrificed when a process gets over-taxed. With disastrous results. Baking communication into the process delivery can free up Associates and ensure that communications is never the chokepoint.

Add Measurement

Improving people-based processes starts with measurement. If you don’t know how the process is working on the ground, you’ll never be able to identify either problems or potential solutions. As with most measurement problems, getting a comprehensive view of performance is complicated and involves multiple types of measurement. Usually, you’ll want to do VoC research on customers to understand whether the process is delivering consistently satisfying results. Tying the VoC to day-time parts can help identify whether there are specific stress times when experiences go south. Similarly, you’ll want to collect operational details about the process. For curbside pickup, that means knowing how many curbside pickup orders are received, the time before they are flagged as ready, the time before the customer arrives, and the number of customer arrivals by day-time part. If available, you’d really like to know the fulfillment time from arrival to completion – but if you’re using a VRU system like the retailer in the example and not collecting the order number, you won’t have that data. Having this VoC and behavioral data will help you identify problem times, but it won’t help you identify problems. To do that, you need to understand the actual work on the ground when problems are occurring. And that means tracking people.

People-tracking is equivalent to session replay in the digital realm. You can pick any time of day, and visually track exactly what happened. You can track one employee or a whole crew. You can track one order or a whole group of orders. And you can see how many employees were working a process, how much time they spent on each step, how the process played out, and how the process changes when loads get high. In short, you get to SEE what actually happened – making both problem identification and process improvement a reality.

Understanding Process Flow and Identifying Loop-Backs

Here’s an example where completing a particular fulfillment required a full re-cycle in the Process:

Many new processes must be sensitive to employee density. Here you can see when processes brought multiple Associates into close proximity:

Studying how processes played out often reveals significant improvement opportunities. In this case, the final step of most pickups required a trip to the back-station. By moving that location near the outside door, 2-3 minutes were shaved off every fulfillment. Analysis also showed several opportunities for order parallelization and identified a problem in the way larger orders were assembled that was significantly slowing down fulfillment. These kinds of improvements simply don’t happen without measurement.

When you make fundamental process improvements, you decrease each customer’s wait time, increase Associate productivity, and reduce the likelihood of process break-downs due to over-demand.

How do you get this kind of measurement?

There are three technologies used for people-process tracking. Each brings specific capabilities to the table that make them suitable for certain use-cases.

1. Camera: Video camera is the most popular edge-based people-tracking solution. Camera solutions provide really high positional resolution and continuous (30 fps) motion tracking of Associates in a work area. With camera, you can know exactly what steps Associates are taking in a complex process and how long each step is taking them. But camera tracking has some drawbacks. It’s the most expensive technology to deploy. It’s inappropriate for really large areas, and it requires continuous line-of-sight to the person. Spaces with lots of blockage – even for ceiling mountain cameras – are bad candidates for camera tracking. It’s also much harder to tie the measurement to a specific individual. So if you need to tie process to individual staff, camera is probably the wrong choice.

2. Bluetooth (BLE) Tagging: Bluetooth tracking relies on Associates carrying small pingers (key-card or belt-clipped thumb-drive sized devices). These devices ping out every second and are detected by Bluetooth sensors distributed around the workspace. Since the sensors are very inexpensive, they can be deployed even in very large spaces. By pinging every second, Bluetooth provides continuous measurement, but it isn’t as positionally accurate as camera. On the other hand, it can work even in visually crowded environments or when people are moving under obstructions or into visually closed off areas (like closets). It can tie at the individual level (or not – if the pingers aren’t specifically assigned). Bluetooth is great for a wide variety of use-cases. It’s only real drawbacks are that it’s not positionally accurate enough for some situations and it can’t provide any information beyond position. Bluetooth doesn’t know if an Associate is sitting or standing, carrying something or interacting with a machine. If you need to know exactly where a person is or capture detailed movement data to understand what they are doing, Bluetooth isn’t a solution.

3. Mobile Devices: You can consider people-tracking using existing mobile devices – phones or tablets – that are already being used and carried. If devices are carried and used for operations, then adding geo-location to them is almost a slam dunk. You’ll need very little additional hardware (adding Bluetooth beacons to increase positional accuracy is a trivial endeavor) and you’ll get pretty good accuracy (slightly better than pingers but not as good as camera). On the other hand, when you add geo-location to mobile devices, you get a direct tie to the operational workflow. So you can see how processes played out in space while following an in-App “paper” trail. This kind of tracking is almost certainly the right solution if your Associates already carry and use devices.

Create Overflow Strategies

Most processes work just fine when they are not stressed. When there’s a single customer order to be processed, using phone for curb-side pickup works. But people processes are inherently constrained and tend to break when they get stressed. So building out overflow strategies for when that happens is critical. With stress points, you shouldn’t think of the entire process. You should dissect each step and how it might behave when stressed. For curb-side pickup, that means breaking down everything from parking to fulfillment to communications. Constrained resources like parking spaces, fulfillment workers and phone operators are particularly likely to be pain points. Designing an overflow strategy for parking involves a combination of pre-messaging (what to do if the spots are filled), on-site signage, and better communications channels. For the process itself, overflow strategies range from better scheduling (don’t overschedule pickups on a Saturday afternoon), real-time communication of wait times so that customers will avoid peak times, better staff allocation models, overflow staffing options, and better communication options for handling Just-in-Time pickup (sort of like Disney’s FastPass) in a very narrow time window.

Far too often, people treat overflow as an act-of-god. They shrug their shoulders as if to say, “We were busy.” But think about the list of potential solutions – from scheduling to process changes – that can be applied. Processes are not helpless in the face of demand. There are nearly always strategies to shift demand, to optimize fulfillment, or to increase fulfillment speed.

If you’ve built measurement into the customer, behavioral and operations of the process, you should be able to see and demonstrate improvement when overflow strategies go into place. Continuous improvement as a discipline was invented for manufacturing processes and it translates remarkably well into this kind of people-process improvement.

Handling Communications

No matter how good your processes become, situations where there’s more work than people to do the work will happen. That means longer wait times and potentially broken customer experiences. That’s why it’s critically important that good customer communications be baked INTO the operational procedures. Part of what made the VRU curbside experience fall apart was that the Associates manning curb-side pickup were too swamped to manage the communications. That’s a common problem in field-based customer operations. People get too busy to communicate – making everything worse. Embedded and automating communication inside the operational processes solves this problem in almost every situation. For curbside, the easiest way to do this is to strip away the VRU and replace it with a mobile texting system. When the shopper arrives, they text their order number. Now they don’t have to wait and nobody has to be on the phone. Their arrival time and order are instantly logged giving you operational measurement. Even better, if employees are given an electronic queue and they swipe orders they’re starting and finishing (which is a heck of a lot easier than answering the phone), then the system can automatically calculate how many orders are ahead of a new arrival. This allows the system to communicate the state of the queue and estimate the wait time. It can do this automatically with NO Associate involvement. By embedding communications inside the process, you ensure that it doesn’t suffer when Associates get super-busy.

This kind of embedded communications also allows for more flexibility in operational improvement. When wait times exceed some threshold, the customer can be offered a specific return time with a JIT pickup. By publishing the current queue on Web and App, you can let customers know what to expect and automatically shift volume into less busy times. Knowledge is power, and accurate customer communication is nearly always preferable even the news isn’t great.

Summing Up

For people-based processes, lack of visibility into how the process works, high-variability in usage creating stress points, and break-downs in customer communication are the biggest challenges. Solving each is critical to driving operational improvement. With continuous improvement based on careful measurement, thoughtful overflow design, and embedded customer communications, you can make almost any people-based process dramatically better.

Measure: Modern people-measurement technologies provide effective avenues for understanding how processes are getting performed. Camera provides very fine-grained movement tracking but is difficult in obstructed environments and isn’t well-suited to tracking specific individuals. BLE and App tracking solve both these problems and are easier to deploy (especially if employee’s already carry devices). But they are not as positionally accurate as camera and cannot track any aspects of movement except position (things like facing, interactions, carries are not trackable with electronics).

Improve: With measurement in hand, improving processes like curbside pickup becomes a lot easier. Measurement lets you drive true improvement in the underlying process execution – improvement which will flow through every aspect of the system.

Overflow Design: Still, even the best systems do get stressed, which is why operational improvement to inherently variable processes like curbside pickup should always address overflow concerns. Figuring out the limitations of each step in the process and what to do when those limits are exceeded will help avoid many of the worst customer experience problems.

Embedded Communications: Embedding communications inside the process is almost always practical in today’s world and it ensures that communications tasks don’t fall by the wayside when people get busy. Since missing or poor communication is often the single biggest driver of customer dissatisfaction, fixing the messaging is a huge win.

Written by: Gary Angel
Reviewed by: Allen Hillery