Type to search

How a data strategist used a digital mindset to save a struggling Vietnamese restaurant

Technology & Data

How a data strategist used a digital mindset to save a struggling Vietnamese restaurant


At Adobe Symposium in Sydney, NAB’s Matthew Peters shares how he took a ‘test and play’ attitude to rehabilitating a local restaurant.

“We’ve seen the evolution of web analytics becoming digital analytics, and now going further and becoming this idea around customer analytics and integrating it with the rest of the data that you have at your organisation,” Matthew Peters told the Future of Enterprise Customer Intelligence breakout session crowd at Adobe Symposium Sydney this morning.

As engineering manager, customer insights and analytics digital content, forms and marketing technology at NAB, Peters runs the customer insights and analytics team within NAB’s marketing function.

In terms of figuring out how to use data to measure what customers are doing, how they’re doing it and improving that experience constantly, Peters thinks “we’ve done a great job.

“Something we think about, and I know this happens at NAB, we think less about ‘who are our bankers?’. 

“Sometimes [considering] the interfaces, tools and number of tools that we have, we haven’t put enough thought into empathy, into what their job is like. If you want to look after your customers, look after your people because they are the ones that are going to look after your customers.”

“You only get value when you make changes. That can be hard for large organisations to do.”

Peters recounts a Vietnamese restaurant in Melbourne’s Mount Waverley that he was quite fond of. Unfortunately, around 18 months ago the restaurant went up in flames as one of the cooks left a pot on the stove overnight. Not only was this restaurant of personal importance to Peters, it was also a customer of NAB’s.

Upon its re-opening, everything appeared to be on the up and up. The new space was bigger and people were flocking back to their local favourite.

“I was having a conversation with the owner and she said ‘we’re just not making money’,” Peters relayed. “I said, ‘what do you mean you’re not making money? Every table, every night is full. You’ve got a line of people out the door, how could you not be making money?’”

So she showed him the restaurant’s end-of-day summary. “I said, ‘let me have a look at the data, not the summary, let me see all the raw data.’”

Peters found that the amount of time between when a customer made an order and paid for their meal was almost two hours, “A really long time, particularly for Vietnamese cuisine which is really fast.”

Upon further analysis, he discovered it was taking an average of 63 minutes for an order to reach the table, “That’s a terrible experience, but thankfully the food was really good so people kind of tolerated it.”

Putting on his data hat, Peters began investigating. “I thought about it: what would I do if this was something at NAB, my day-to-day job? I would say, ‘we need to change this, let’s test this, let’s go to test and targets’. Well why can’t I do that for a point of sale (POS) system at a restaurant? It’s something I’ve been doing with Adobe tools for the past 10 years.”

Walking through the kitchen looking for clues, something Peters likened to Gordon Ramsey’s Kitchen Nightmares, he found one of the restaurant’s service dockets – the slip of paper passed from floor to kitchen staff that was supposed to communicate order details. It was a mess, nearly impossible to tell how many items the order contained, which items were entrees or mains; key information wasn’t prominent enough and unimportant information was taking up too much space.

So Peters did what any good engineering manager would do and reformed the order slip’s UX. He made the table number the focus of the slip, ensured that all items spanned fewer than 19 characters (so not to take up more than one line) and separated items into entrees and mains.

“Think about the employee experience. Everything you can do to make their lives easier is an improvement,” said Peters.

After only a few days of implementing the new slips, food was no longer being delivered to the wrong tables and the time it took for orders to reach customers had halved. 

But Peters wasn’t done. Turnover had improved, but tables were always full, “it seemed like we should have had more free tables.” He discovered another bottleneck, it was taking too long for service staff to enter orders into the POS system. Dockets would be piled up next to the counter while staff were busy helping other customers.

“This feels like what a lot of our bankers at NAB go through everyday. Systems that are hard to use and are slow.” Peters stressed the importance of having a testing mindset, approaching issues with a data-first attitude

“Everything we can do to make our systems and employee experience better is going to be an improvement.”

The POS was clunky and unintuitive, employees were finding it very difficult to efficiently communicate order details through the interface. So Peters did another overhaul, colour coding, adding tags such as ‘extra meat’, ‘no peanuts’, ‘vegetarian’ or ‘supersize me’ so employees wouldn’t have to waste time typing out notes. Essentially, he wanted to make the system as easy to use as possible. “You want the guys to just go tap, tap, tap and that’s it, order done. That saves a lot of time.” As a result, the time it took to place an order went from three or four minutes to 30 seconds.

As a result of Peter’s data-focused reform, the restaurant’s service time went down from 63 minutes to a staggering six minutes.

There was just one more element that needed a little attention: the menu.

“We’ve looked after the employee experience and tried to make their lives easier. What else can we do around customers are doing? I looked at the menu and I thought, ‘this menu is just like any other web page that I’ve had before.’ Why can’t you do optimisation on a menu? Why can’t you do testing? 

“You print them once and they seem to stay there for six months. That’s not how we do it in digital. You could print a new menu every couple of days and try it – give half the restaurant one menu and give half the restaurant another menu and see what the results are. That’s what we did.”

Essentially, Peters was A/B testing physical menus with live customer samples.

The original menu was printed in A3 size, featuring a relatively standard layout for casual Asian restaurants. One lesson of ‘menu engineering’ Peters employed was the placement of prices next to items. “Any time you have a currency sign or your prices listed in a column, what people do is look down the list and they tend to shop by price,” Peters explained.

Example menu – prices column

Example menu – prices in a column

“You’ll see this [instead] at fancy restaurants, the price is then interlaced without a currency symbol, without decimal places. That really helped.”

Example menu – prices interlaced with items

Example menu – prices interlaced with items

After six weeks of testing iterations, the menu had reduced in size, prices were listed in whole numbers in item descriptions and the restaurant’s pho dish – which made up 23% of sales – was made more prominent as a ‘hero’ item. “That’s what you want, for people to order lots of the thing that’s really easy, cheap and fast to make. It’s also delicious and makes you money.” Another concept of menu engineering.

“Next time you’re eating out somewhere, look at the menu and see if you can pick out these things. They’ll ‘hero’ items or dishes, probably the ones they’re making a lot of money on.”

In the end, Peters had made a total of 38 changes to the restaurant’s operations. Sales increased by 41%, service time reduced by 90.5%, review sites were rating the restaurant between four and a half and five stars and TripAdvisor had even listed it as the number one restaurant in the area.

It turned out that the manufacturer of the POS software (which previously had been slowing down the business) heard about how Peters had transformed the restaurant. It reached out to ask if it could use the story as a case study, demonstrating how the POS could do the same for other service establishments.

“I thought, no. This is the point. It’s not the software,” Peters affirmed. “You can have the greatest software, if you’re in this room [at Adobe Symposium], you’re probably using Adobe products, you’ve got the best stuff. It’s how you use it, it’s how you are data-driven in what you do and consider things.

Peters put to the room, “Ask yourself: are you using your data as a thermostat or a thermometer? Are you using it as something that you’re always changing and seeing what happens? Or are you using it just as something just to report and get measurements?

“How do we change our whole engineering structure and the way that we develop software in our products so that we can be truly fast in making product changes? We’re now in an era where we have data coming in very quickly. Unless you have that insight, unless you can act on it, it’s not going to be very useful or valuable to you.

“In every organisation, yes, customer experience is very important. Customers of this [restaurant], which is a customer of NAB, are very happy. But the employees are happy as well. So when you’re bringing data in and you’re thinking about customer intelligence and the data you’re collecting, have a think about employee experience. That’s certainly something I’m thinking about.”


Further Reading:




Full disclosure: the author of this article is attending Adobe Symposium as a guest of Adobe.





Image credit:Markus Winkler

Josh Loh

Josh Loh is assistant editor at MarketingMag.com.au

  • 1

You Might also Like

Leave a Comment