Summary. Store operations wants mobile training because it keeps associates on the floor. Finance wants a number. This paper sets out the four places the return actually comes from, which parts are arithmetic from your own figures, which parts rest on published research, how to answer the attribution objection, and which figures we deliberately refuse to count. Every external number is footnoted.

The question worth arguing about

Whether a device belongs in an associate’s hands is settled. No retailer needs persuading that store staff should be able to look something up.

The live question is narrower: can the person on your floor tell a customer something the customer did not already find on their phone? That is not a technology question. It is a question about what the associate knows at the moment of the conversation, and it is the only thing a store can offer that a product page cannot.

So the case for on-floor training no longer rests on novelty. It rests on arithmetic, which is what follows.

The argument that loses in the finance meeting

The case for mobile training usually arrives as a list of benefits. Associates stay on the selling floor. Content reaches every store the same day. Managers can see who engaged. New hires ramp faster.

All of it is true, and none of it survives contact with a CFO, because every item is a capability rather than a cash flow. The question being asked is fairer than it feels: what changes on the P&L, and how would we know?

There are four honest answers. They are worth separating because they carry very different burdens of proof.

Driver one: selling time returned to the floor

This is pure arithmetic from figures you already hold. Take the training hours you currently deliver off the floor. Move the share that can be delivered in between customers. Multiply by your own revenue per associate hour, and apply your own gross margin.

Nothing here is an estimate about human behavior. It is a scheduling change expressed in money. The only judgment call is what share of your training hours can move, and you are better placed to answer that than any vendor.

The same change produces a cost saving alongside the revenue: hours moved out of dedicated sessions stop consuming scheduled payroll, stop requiring a trainer, and stop requiring travel.

Driver two: full-price sell-through

Markdown reduction is usually filed under cost control. It belongs on the top line. Selling a unit at full price rather than thirty percent off raises average unit retail, which raises revenue on the same unit volume.

This is also the one lift driver where we hold a measured client result. A specialty apparel retailer moved its markdown rate from 18.1 percent of sales to 14.9 percent across a full year following mandatory product knowledge training, a 17.7 percent relative reduction against a plan that called for the rate to rise.⁴

In our own calculator we default this driver to roughly half that result rather than the full figure. One client, however well documented, is not a fleet.

Driver three: sales per hour from better-trained associates

This is the driver most often oversold, so precision matters.

Marshall L. Fisher, professor of operations, information and decisions at the Wharton School, studied a large department store chain with roughly 300,000 sales associates over three years, alongside millions of customer satisfaction responses. Associates who received any training averaged 46 percent more sales per hour than those who did not.³

The number to quote is not 46 percent. The researchers split that gap, attributing half to the individuals and half to the training itself, which leaves roughly 23 percent as the training-attributable figure. We use 23 percent, because 46 percent invites the obvious objection that your trained associates were already your better associates, and Fisher has already netted that out.

Because the finding is expressed in sales per hour, it composes directly with the revenue per associate hour you already calculate. No new conversion rate assumption is required.

Driver four: turnover you stop paying for

This is the driver most often left out of a training business case, and on current numbers it is frequently the largest single line for a big fleet.

McKinsey puts the cost of losing one frontline retail employee at nearly $10,000 on average, counting wage levels, covering the vacant shifts, the time to hire a replacement, and the time that replacement takes to reach peak performance.² Against that, Korn Ferry’s survey of more than 100 major US retailers found hourly in-store turnover at 75.8 percent, up from 68 percent the prior year, with part-time hourly running near 85 percent.¹

The reason this belongs in a training business case rather than an HR one is what the same survey found about cause. The top reason people leave a retail job is for a better opportunity or promotion, and Korn Ferry’s recommendation is explicitly that retailers provide clear advancement paths and training.¹ Training is not only a productivity lever. It is a retention lever, and retention is priced at roughly $10,000 a head.

Multiply your associate count by your turnover rate, by whatever share of that turnover you believe training influences, by your own cost per hire. Keep the third number conservative and the line still lands hard.

“None of this is causal”

This is the objection a good finance team will make, and it is correct about most published training research. It should be answered with a method rather than with more citations.

Two of the four drivers do not need causal evidence at all. Selling time returned to the floor is a scheduling change, and the print and payroll savings are invoices you stop paying. Those are arithmetic, and they can carry the business case on their own if the lift drivers are disallowed entirely.

The two lift drivers can be tested in a quarter. Hold a matched set of stores back. Comparable format, comparable volume, comparable market. Deliver to the rest. Compare the two groups over eight to twelve weeks, rather than comparing before against after, so the promotional calendar and the season cannot explain the result.

Run it on the driver with the cleanest signal. Markdown rate and return rate are better test candidates than units per transaction, because they move for fewer reasons and are recorded consistently.

Agree the failing condition in advance. Write down, before launch, the result that would cause you to stop. A test with no failing condition is a rollout with extra paperwork, and finance can tell the difference.

The honest position: published research establishes that the effect exists and roughly how large it can be. Only your own held-back comparison establishes that it exists in your business, and it costs one quarter and no additional software to find out.

What our clients have reported

• Print and shipping eliminated. One retailer with 400 locations reported saving more than $500,000 a year when digital content replaced the printing and shipping of its quarterly knowledge cards.
• Basket size. A global accessories and lifestyle brand recorded a $20 increase in average dollars per transaction at one store, and a 21 percent lift in handbag sales at another in a single quarter. Both are individual store results rather than chain averages, and the write-up says so.⁵
• Scheduled training removed. The same brand eliminated a monthly two-hour training session, along with the payroll it consumed.

A worked model, labeled as a model

For a retailer with $400 million in revenue, 5,000 associates and 400 locations, the drivers above produce a figure in the range of $13 million to $14 million of annual value, of which roughly $3.8 million is hard cost savings and the remainder is margin on influenced revenue.

Treat that as an illustration of the arithmetic, not a promise. It is built from a representative set of inputs and moves substantially when they move. The version that matters is the one built from yours.

What we deliberately do not count

Our calculator counts margin, never revenue, on every influenced dollar. It excludes the best individual store results we hold, including the 21 percent handbag lift and a 70 percent lift on a single accessory item after one training issue, because a figure from one item or one store does not survive being applied to a fleet. It values on-the-floor training only, and puts no number on tasks, forms, audits or store communications.

Stating what we refuse to count is not modesty. It is the reason the remaining numbers are worth reading.

Taking it to the meeting

Four questions come up every time, and having the answer ready is most of the work.

Which line does this hit, and when? Payroll and print savings land in the first quarter. Markdown and sales-per-hour effects need a season. Say so explicitly rather than blending them into one annual figure.

What if adoption is poor? Show the model at half your assumed engagement rate. If it still clears the cost, the decision does not depend on optimism, and that is a much stronger position than a single confident number.

Why now rather than next year? The turnover line is the answer. Every year of delay is another full cycle of replacement hiring at your own cost per hire, absorbing the capability you already paid to build.

How will we know it worked? Name the held-back stores, the metric and the date, before approval rather than after. A proposal that arrives with its own test design is a different kind of proposal.

Run it on your own figures

Every driver above is expressed in inputs you already have: revenue, associate count, paid hours, training hours, gross margin, markdown rate, turnover rate, cost per hire. The calculator assembles them, shows its working, and emails you a spreadsheet in which every result is a live formula rather than a static snapshot, so your finance team can change an assumption and watch the model move.

Build the model on your own numbers with the ROI calculator.

References

1. Korn Ferry, Retail Employee Turnover on the Rise, November 15, 2022. Survey of more than 100 major US retailers with annual revenues from $100 million to more than $20 billion; fieldwork early November 2022. kornferry.com
2. McKinsey & Company, How retailers can build and retain a strong frontline workforce in 2024, 2024. mckinsey.com
3. Marshall L. Fisher, Wharton School, University of Pennsylvania. Why Training Your Employees Will Increase Sales, Knowledge@Wharton, September 11, 2015. Roughly 300,000 sales associates over three years. knowledge.wharton.upenn.edu
4. Multimedia Plus client case study, reported by the client from fiscal year-end reporting. Retailer not identified. Markdown rate down 17.7% year over year
5. Multimedia Plus client case study, figures stated publicly by the brand’s retail marketing leadership at NRF’s Big Show, January 2017. Retailer not identified. $20 more per transaction after training moved to the sales floor