ngunji.com (author’s own site) · Author’s notes on an interview

What Makes Retail DX Pay Off: Notes on My Interview with DXportal

DXportal interviewed Noboru Gunji for its “Ask the Experts” series. In this post he sets out the main points: start from customer value, not technology; watch for departmental sub-optimization; and make the desired state concrete.

English translation of the author’s original post

Author
Noboru Gunji (郡司 昇)
Original title
小売DXで成果を出すには?利益につながる変革の本質をDXportalで語りました
Originally published
, in Japanese, on the author’s own site
Original last updated
English version

My interview has been published in DXportal’s “Ask the Experts” series. The theme is what lies at the heart of “transformation that produces results” in retail DX, and the future that executives should be picturing.

From working in stores to running drugstores, integrating systems and logistics, bringing an e-commerce business into the black and driving an omnichannel strategy: drawing on that practical experience, I talked about how to connect the introduction of digital tools to customer value and profit.

The starting point of retail DX is not the technology you install. It is the value your company wants to provide to customers, and the desired state that would deliver it.

In this post I introduce the main points of the interview and add some perspectives for reviewing your own company’s efforts.

The interview

  • Publication: DXportal
  • Published: August 12, 2026
  • Updated: October 5, 2026

The purpose of retail DX is customer value and profit, not installing digital tools

Introducing AI or data analysis tools does not, by itself, achieve DX.

The “X” in DX is transformation. Digital technology has to be used to change operations, decision-making and the way value reaches the customer.

In the interview I divided the directions for raising customer value into two.

  • Remove the frustrations of shopping. Reduce inconveniences such as the product not being there, not being able to find it, and waiting at the register.
  • Add reasons to choose the store. Raise the quality of service and solve customers’ problems better.

What matters is connecting these changes to business results.

For example, even if you shorten the time spent on ordering, it is not a success if stockouts increase and sales opportunities are lost. Even if you introduce AI to support customer service, the introduction has no value if customers do not get the explanations they need.

Evaluation should rest on “what changed for the customer, and how did it lead to profit,” not on “what did we install.”

What blocks results is sub-optimization by department

If the challenges of retail DX are treated as a problem for the information systems department alone, the scope of what needs solving will be misjudged.

Merchandising, logistics, store operations, e-commerce, marketing: each of these is connected to the others. An improvement in one department can increase the burden on another.

The example I gave in the interview was purchasing.

A buyer prioritizes earning rebates and buys large quantities of products that do not sell. By the merchandising department’s measures it is a success, but in the stores the burden of inventory, stocking shelves and markdowns grows. Looking at purchase terms alone will not tell you whether profit increased for the company as a whole.

When advancing DX, at least the following three points should be checked.

  • How the work connects. Including the steps before and after, where do rework and bottlenecks occur?
  • Whether responsibility and authority match. Can the person driving the project decide the changes that are needed?
  • Whether the metrics line up. Do a department’s results lead to customer value and to profit for the whole company?

Review not just the choice of tools but the workflow and the system of evaluation. That is the condition for moving from partial to overall optimization.

Read success stories for their preconditions, not their methods

In the interview I also touched on cases such as the home-improvement chain Goodday and UNIQLO overseas.

But using the same tools as another company does not produce the same results.

If the trade area, customers’ shopping behavior, store size, staffing and logistics conditions differ, the right measures differ too. Take in-store pickup: at a store where most customers arrive by car and one where most arrive on foot, the handover location and the way it is run will be designed differently.

When learning from a case, there are three things to check.

  • Which customer problem was the company trying to solve?
  • What operational and organizational conditions supported the result?
  • Which parts can we use, and which do we need to change?

Success stories are for building your own hypotheses, not for borrowing answers.

Make the desired state concrete enough to check on the shop floor

What I conveyed at the end of the interview was the importance of executives making “the state we ought to be in” concrete.

“Strengthen customer service” or “use AI” does not tell people on the shop floor what to change.

If customer service is to be a strength, for example, make it as concrete as: “a state in which even a new staff member can check the information a customer needs to choose a product and explain it with reasons.”

Working backward from there, the necessary measures can be broken down: organizing product information, procedures for handling inquiries, training, support from AI, and so on.

A table like the following helps in organizing the effort.

Customer problem Desired state Examples of metrics to check
Cannot find the product they want Can check the product’s location and stock without confusion Time spent searching for products; accuracy of stock guidance
Kept waiting for an answer to a question Staff can check the necessary information and explain Time to answer; rate of repeat inquiries
In-store pickup takes time Can receive the order accurately and without waiting Handover time; rate of wrong shipments

This is a design example added for this post. Metrics need to be set to fit each company’s circumstances.

Alongside customer-experience metrics, check sales, gross profit and operating costs. For work hours that have been cut, it is desirable to follow through to what that time was redirected to and what result it led to.

Use firsthand information from the field as the basis for transformation decisions

I make a point of traveling abroad three or four times a year and aim to see 500 stores a year, including those in Japan.

The purpose is not only to look for new equipment. It is to check how customers shop, how employees work, and how the systems are actually run.

A service that looks convenient in the announcement materials can turn out to be a hassle when you use it. Conversely, an inconspicuous change in operations can be what makes shopping easy.

It is precisely because AI has made information easier to gather that checking in the field has value. In retail DX, connecting information about technology with the reality of the stores is indispensable to good judgment.

About the full interview

In addition to these points, the DXportal interview covers my experience bringing an e-commerce business into the black, my book The Essence of Retail 2025 DX (小売業の本質2025DX), and Latest Trends in Retail and Stores 2026 (小売業と店舗の最新トレンド2026), which I co-supervised.

I hope it will be read by retail executives and those in charge of DX, and also by people at IT companies developing services for retailers.

← All articles