by W H Inmon, Forest Rim Technology

The restaurant business is competitive. Restaurants are always competing for new customers in many ways. And restaurants want to hold on to their existing customer base.

Typically restaurants are measured by several parameters:

  • Menu
  • Ambience
  • Service
  • Pricing

While there are other aspects to managing a restaurant, these measurements are the most commonly used and are the most important.

The Voice of the Customer

But success in managing a restaurant always boils down to one thing – the opinion of the customer. The restaurant can rate itself all it wants in whatever categories it wants. But the only rating and opinion that really counts is that of the customer.

For long term success and health the restaurant needs to listen – continuously – to what the customer is saying about the restaurant. It is not enough for the restaurant to hear the voice of the customer today. The restaurant must hear the voice of the customer tomorrow, next week, and next month.

There are many ways feedback to the restaurant can be accomplished. There is twitter. There are paper feedback forms placed at every table.

But in today’s world, feedback is often achieved by allowing the customer to voice his/her concerns about the restaurant over the Internet.

Feedback over the Internet

The Internet is a good medium for hearing what the customer has to say for several reasons. Over the Internet, the customer can be anonymous if the customer wishes. Over the Internet the customer can be as explicit as the customer desires. And over the Internet the customer can write as long or short a message as the customer wishes.

For these reasons and more the Internet is an ideal mechanism for the exchange of feedback from the customer to the restaurant.

An Example

As a simple example of the feedback given to a restaurant, consider the following (real) example –

Over the course of a month the restaurant will receive MANY message similar to this one.

There are many lessons to be learned from looking at customer feedback. But there are some problems associated with that feedback.

Practical Issues

The first problem is that there are so many messages from the customer. In a month’s time there may 50,000 to 100,000 messages for a good sized restaurant chain. Trying to manually read and digest 50,000 messages manually is an impossible task.

One restaurant chain prints the monthly messages out and places them on the desk of the appropriate vice president. The messages sit there on the desk until next month when another 50,000 messages are printed out.

No one even looks at what the customers are saying.

So the volume of messages is one major problem.

The second major problem is that the messages are in text. The computer does not handle text well. The computer is designed to handle repetitive, transaction activities, where there is a high degree of structure to the activities. And there just is no high degree of repetitivity to be found in text.

These then are some of the reasons why customer feedback is so hard to turn into useful information.

TEXTUALETL

Now there is technology that greatly alleviates the challenges facing the organization trying to hear the voice of the customer.

That technology is Textual ETL by Forest Rim Technology.

With Forest Rim Technology now – for the first time – you can easily and accurately hear the voice of the customer. Now the restaurant can actually listen to ALL of their customers in an accurate, easy to understand manner.

The architectural rendition of how Textual ETL works looks like –

Textual data – customer comments – are fed into Textual ETL. You can feed as many as you like into Textual ETL and the comments that are fed in are in text – narrative. Textual ETL then reads, analyzes, and converts the comments into a data base. Once the comments have been converted into a data base, they can then be analyzed by a standard analytical program. Then management can see what the customer is saying.

So…what is the Customer Saying?

As an example of the analysis that can be done, consider the following –

The graph shown has seven categories – other ingredients, people, place, price, process, product and promotions. The graph was created from an analysis by textual ETL of thousands of customer comments.

In the graph red indicates a negative comment and green indoicates a positive experience. At first glance it appears that there is a lot more red than green, which is true.

But industry experience has shown that under normal circumstances people only give feedback when they are dissatisfied. When a person goes to a restaurant and has a positive experience, no feedback is given. The expected ratio of bad to good experiences is 85% to 15%. The the ratio is different from this then something may really be going wrong.

So the fact that there are more red than green is not a great concern.

There is one really remarkable fact that sticks out from this chart. That fact is no one says anything about price. This is an indication that the restaurant chain is “leaving money on the table”. The restaurant chain needs to consider marginally raising prices.

Another interesting fact that comes from this graph is that hardly anyone has anything to say about promotions. In fact the restaurant chain is doing promotions. But the promotions are having little or no impact. The restaurant ought to consider doing some other kinds of promotions.

Giving management the message that they ought to be charging more and that they ought to be doing more effective promotions is important news that management ought to hear.

And where did this information come from? From the customer himself/herself.

But there are lots of other pieces of information that can be gleaned from what the customer is saying.

Consider the graph –

In this graph the different dishes in the pasta category that were mentioned by the customer are shown.

The most mentioned dish was penne rosa and the least mentioned dish was Wisconsin macaroni and cheese.

This graph then shows what was on the mind of the customer after having a dining experience.

But there is other important information that can be gleaned. If you – as a manager are interested – you can drill down on any given dish. For example you could take a closer look at Pad Thai –

When you drill down on Pad Thai you see that most of the comments were expressing dissatisfaction. Continuing your curiosity you ask – WHY? – are people dissatisfied with Pad Thai? Is Pad Thai too spicy? Not spicy enough? Too many noodles? Not enough sauce? What is it about Pad Thai that the customers are not liking?

The further drill down shows that there are positive reasons for the customers reactions.

On the negative side it is seen that portion size is the main reason for the customers negative comments to Pad Thai. The restaurant needs to increase the portion size of Pad Thai in order to please the customer.

Not only is sentiment derived from the customers comments, but the reasons for the sentiment is also derived. This is really powerful information for the manager that wants to increase customer satisfaction.

But there are other powerful pieces of information that can be gathered from the customers feedback.

Another way to look at information is by information over time. Once the data has been put into a data base, it can be analyzed over time.

Another possibility is looking at customer feedback based on the stores (the physical locations where the the restaurant chain has businesses) –

In this figure it is seen that the different locations are ranked by comments per location.

In this case it is seen that store 221 has had more feedback than any other store.

This may or may not be an indicator of a peroblem at store 221. It could be that store 221 is in mid town Manhattan and does more business than any other store. Or it could be that there is a real problem with store 221. If this information were married with other corporate information you could find out whether this is a problem location or whether this location is just large.

In any case, management knowing which locations are well run and which locations need management attention is VERY VALUABLE INFORMATION.

And where did this information come from? From the customer himself/herself.

Suppose now management wants to find out what exactly is going on in location 221. With the data in a data base, they can do exactly that.

The graph shows that with the data in the form of a data base, management can look deeply into what is going on in location 221.

In Summary

When you are running a restaurant, hearing what your customers are saying is the key to return business and a solid foundation of loyalty. With the Internet it is now possible to have a direct connection between you and your customers.

And with Textual ETL you can put the feedback from the customers in a form that can be analyzed. With the Internet and textual ETL you can –

  • Listen to thousands of customers
  • Understand text and narrative

Once your data is in the form of a data base, you can analyze it many different ways. You can find out sentiment. You can investigate sentiment to find out why sentiment is the way that it is. You can drill down on any issue you wish.

You can find out what locations have the most complaints. You can find out what is behind the complaints.

You can investigate important issues such as – are we charging too little? How effective are our promotions?


Forest Rim Technology was formed by Bill Inmon in order to provide technology to bridge the gap between structured and unstructured data. Forest Rim Technology is located in Castle Rock, Colorado.