Calculate LMS ROI before you buy by building a line-item table, not by borrowing a percentage: price every hour of staff time at a loaded rate, keep the cash you’ll stop spending apart from the hours you’ll free up, and leave performance gains out of the number until you can show the platform caused them. Your director will test the figure one line at a time, and a return that can’t say where each dollar comes from won’t survive the first “how do you know”. This guide builds the table for a 400-person company, shows why its first year comes out negative and its payback lands in month 16, and sets the rules that make each line defensible.
Key Takeaways
- Price staff time at salary ÷ 2,080 hours × about 1.43. Benefits made up 30.0% of private-industry compensation costs in June 2026, per the Bureau of Labor Statistics.
- Split benefits into cash (spend that stops) and capacity (hours that come free). In the worked example, the cash lines never cover the subscription on their own.
- One purchase, two answers: -13.4% ROI in year one, 64.3% in year two. Report the payback month (16 in the example) beside both.
- A turnover or productivity gain enters the number only with a method that isolates the platform’s share: a comparison group, a trend line or a confidence-adjusted estimate.
- “$11,000 saved annually” with no cost beside it isn’t a return you can compare.
Course building is the biggest one-time cost in the table below, so time one real course in an AI LMS during the trial and put the hours you measure into that row.
The LMS ROI Formula, and the Model Behind It
The formula is short: subtract total costs from total benefits, divide by total costs, and multiply by 100. It comes from the Phillips ROI Methodology, developed by Jack and Patti Phillips of ROI Institute, which writes it as ROI (%) = Net Program Benefits ÷ Program Costs × 100. ROI Institute pairs it with a benefit-cost ratio (benefits divided by costs) and a payback period.
The model measures five levels: Reaction and Planned Action, Learning, Application and Implementation, Business Impact, and ROI. Only the last is a money figure. Before you buy, the first four haven’t happened yet, so a business case is a forecast of level five built from lines you can price today.
The formula isn’t where business cases fail. The inputs are. Three rules from the same model decide what goes into the table:
- Load every cost. ROI Institute’s cost categories include software, design and development, administration, and the salaries plus benefits of participants for the time they spend in the program.
- Count saved time only if it goes somewhere. Price it at wages plus benefits, which the application guide calls “a standard formula in most organizations.” Then check the condition that comes with it: “The time saved must be legitimate, where the time savings is used on other productive work.”
- Leave out what you can’t convert credibly. Benefits you choose not to put a money value on are listed as intangibles beside the number, not inside it. The last test before converting a measure is whether the conversion can be explained to an executive “and secure buy-in within two minutes.”
Hold every line to that last test. If you can’t explain a row to your director in two minutes, it goes under the table, not in it.
Price an Hour of Staff Time First
Most lines in an LMS business case are hours. Your team’s hours setting the platform up, your hours no longer spent chasing completions, new hires’ hours no longer spent in a classroom. So the hourly rate you choose moves nearly every row at once, and it’s the first number your director will question.
Use the loaded rate:
- Divide annual salary by 2,080. That’s 40 hours a week for 52 weeks. An $80,000 salary is $38.46 an hour.
- Add benefits. In June 2026, wages were 70.0% of employer compensation costs for private-industry workers and benefits were 30.0%, according to the Bureau of Labor Statistics. Total compensation is wages divided by 0.70, so multiply by about 1.43. The $38.46 becomes $55.00.
- Use finance’s figure if there is one. Finance knows your real benefits bill. BLS also counts paid leave as a benefit, and 2,080 hours already includes leave, so the national multiplier runs a little high on a salary basis.
The alternative is opportunity cost: what the person would have produced in that hour, such as a sales rep’s revenue per hour. It’s the wrong rate for this table. It assumes the saved hour turns into output, which is a claim about cause, and that’s exactly what the “used on other productive work” rule asks you to show first.
The rate prices hours. It doesn’t turn them into money. Keep two kinds of benefit apart from here on:
- Cash lines are spend that stops: an invoice you won’t pay, a tool you’ll cancel, travel you won’t book.
- Capacity lines are hours that come free. They’re real, but they only become money if the work they’re freed for exists.
Directors read these two very differently, and your table should let them.
The Worked Example: An LMS ROI Table for a 400-Person Company
The example is a 400-person company with an L&D team of two. Today, new hires spend a classroom day on onboarding, compliance workshops are bought from an outside provider, a separate quiz tool is paid monthly, and completions are chased in a spreadsheet.
Every figure below is an assumption or a calculation from one, and it’s labeled that way. Replace them with yours. None of them is an industry average.
The inputs
| Input | Value | Basis |
|---|---|---|
| L&D staff salary | $80,000 | Assumption |
| L&D loaded hourly rate | $55.00 | Calculated: $80,000 ÷ 2,080 × 1.43 |
| New-hire salary | $50,000 | Assumption |
| New-hire loaded hourly rate | $34.38, used as $34 | Calculated: $50,000 ÷ 2,080 × 1.43 |
| New hires per month | 10 | Assumption: your hiring plan |
| Setup before go-live | 2 months | Assumption |
Year-one costs
| Cost line | Calculation | Year one | Basis |
|---|---|---|---|
| Subscription | $2,250 a month × 12 | $27,000 | Assumption: replace with your quote |
| Implementation fee | 10% of the first-year subscription | $2,700 | Assumption: ask each vendor |
| Setup and migration | 120 L&D hours × $55 | $6,600 | Assumption |
| Converting existing courses | 10 courses × 20 hours × $55 | $11,000 | Assumption: time your last course build |
| Running the platform | 12 L&D hours a month × 12 × $55 | $7,920 | Assumption |
| Total | $55,220 |
The subscription is the one row you shouldn’t estimate. Use the quote on your own learner count, and check it against what LMS platforms publish at different headcounts. The 10% fee follows a published structure: Continu prices implementation at “10% of the first-year contract value”. Setup is the row teams undercount, and an LMS implementation plan breaks those hours into tasks with owners.
Year-one benefits
Benefits start at go-live in month three, so year one carries 10 months of them.
| Benefit line | Type | Per month | Year one (10 months) | Basis |
|---|---|---|---|---|
| Outside compliance workshops no longer bought | Cash | $800 | $8,000 | Assumption: last year’s invoices, $9,600 |
| Quiz tool cancelled | Cash | $300 | $3,000 | Assumption: current subscription |
| Completion chasing no longer done by hand | Capacity | 24 hours × $55 = $1,320 | $13,200 | Assumption: log one typical month |
| Classroom hours replaced for new hires | Capacity | 10 hires × 5 hours × $34 = $1,700 | $17,000 | Assumption: an 8-hour classroom day becomes 3 hours online |
| Trainer delivery hours freed | Capacity | 12 hours × $55 = $660 | $6,600 | Assumption |
| Total | $4,780 | $47,800 |
Not counted: new hires reaching full output sooner, fewer compliance errors, lower first-year turnover. They sit under the table as intangibles until a method shows the platform’s share.
What the table says
| Result | All lines | Cash lines only |
|---|---|---|
| Year-one costs | $55,220 | $55,220 |
| Year-one benefits | $47,800 | $11,000 |
| Year-one net | -$7,420 | -$44,220 |
| Benefit-cost ratio | 0.87 | 0.20 |
| Year-one ROI | -13.4% | -80.1% |
Two things stand out. Year one is negative because $20,300 of its cost is one-time: the implementation fee, setup and course conversion. That’s normal for a platform purchase, and it’s the reason to report more than one year.
The cash lines never pay for the platform. They bring in $1,100 a month against $2,910 a month in subscription and running time. This business case rests on capacity, so the director’s first question will be where the freed hours go.
Write the answer into the table: the second trainer you won’t need as hiring grows, or the backlog of courses the team can finally build. With no answer, the capacity lines are a weaker argument than they look.
One row deserves a sensitivity line of its own. Course conversion is the largest one-time cost, and AI drafting targets exactly that work. Mini Course Generator says its AI Course Creator “cuts instructional design time by up to 90%.”
Treat any vendor’s “up to” as a ceiling, not a planning figure. At that ceiling, the conversion row falls from 200 hours to 20, year-one costs drop to $45,320, and year-one ROI turns positive at 5.5%. The version of that row you can defend is the one you time yourself on a real course during the trial.
Payback Month and the Second Year
A single percentage hides the shape of the purchase. The same table gives -13.4% in year one and 64.3% in year two, and neither tells your director when the money comes back. The payback month does.
Count it month by month. Months one and two carry the $20,300 of one-time costs plus two months of subscription and running time, so the running total sits at -$26,120 at go-live. From month three, each month adds $4,780 in benefits against $2,910 in costs, a net gain of $1,870.
| End of month | Running net |
|---|---|
| 2 (go-live) | -$26,120 |
| 6 | -$18,640 |
| 12 | -$7,420 |
| 16 | +$60 |
| 24 | +$15,020 |
The platform pays for itself in month 16. ROI Institute’s payback formula divides program costs by program benefits, which works when benefits start on day one. With a two-month setup and one-time costs up front, the monthly count is the honest version.
| Year one | Year two | Two years | |
|---|---|---|---|
| Costs | $55,220 | $34,920 | $90,140 |
| Benefits | $47,800 | $57,360 | $105,160 |
| Net | -$7,420 | $22,440 | $15,020 |
| ROI | -13.4% | 64.3% | 16.7% |
ROI Institute’s guiding principles count only the first year of benefits for short-term solutions. A platform bought on a multi-year contract isn’t one, so show the term you’ll sign. Show each year separately, though, so a strong year two can’t hide a weak year one.
Year two isn’t year one repeated. Before you copy a benefit forward, check what it depends on:
- The renewal increase. An 8% increase at renewal puts the subscription at $29,160 and year-two ROI at 54.7%. That increase is set by one of the four clauses to read before you sign.
- The hiring plan. The new-hire line assumes 10 hires a month. If hiring halves, that line drops to $850 a month and year-two ROI falls to 35.1%.
- Work that comes back. Course conversion doesn’t recur, but updates do when a policy changes. In this example they’re inside the 12 monthly running hours. If yours aren’t, add a row.
When a Benefit Needs Proof: Turnover, Productivity and Errors
The lines in the worked example need checking, not proving. When you stop buying a workshop, the invoice stops. When the classroom day isn’t held, the hours are back. You can confirm each one after go-live against a receipt or a calendar.
Performance benefits are different. New hires reaching full output sooner, fewer compliance errors, lower first-year turnover: each of those numbers also moves when a manager changes, pay changes or the busy season starts.
Before you buy, you have no data on any of them. ROI Institute’s guiding principles cover that case: “If no improvement data are available for a population or from a specific source, assume that little or no improvement has occurred.” So they stay out of the base case.
They don’t disappear. List each one under the table with the metric named, and plan how you’ll isolate the platform’s share, since the same principles call for “at least one method to isolate the effects of a program”. Three are within reach of a two-person team:
- A comparison group. Run the pilot with one team. Keep a similar team on the old process for the same months. With 10 to 20 people in the pilot, treat the gap between them as a directional read, not proof.
- A trend line. Plot the metric for the months before rollout and project it forward. The gap between that projection and what actually happens is the program’s share, provided nothing else new entered the picture.
- A confidence-adjusted estimate. Ask the people who took the training two things: what share of their improvement came from it, and how sure they are. Multiply the two. In ROI Institute’s own example of these participant estimates, a factor credited with 60% of an improvement at 80% confidence is claimed at 48%.
That gives you the answer when your director asks about turnover: it’s listed, it isn’t counted, and here’s how you’ll know. Once the platform is running, reporting those numbers is its own job, covered in which training numbers you can report and defend.
Reading a Vendor’s ROI Figure
Vendors will offer you a number, and it’s tempting to borrow one. Two examples show why it can’t go into your table.
Absorb leads its ROI calculator with “490% ROI Achieved On Enterprise Learning,” from a Forrester Total Economic Impact study that Absorb commissioned. The calculator asks only which use cases you have, then returns a dollar total “based on real customer input found in the TEI Study” plus industry data. That’s a figure about other organizations, built from lines you can’t see.
Docebo describes MidFirst Bank’s result as “a positive ROI of $11,000 saved annually,” without saying what the program or platform cost. A saving with no cost beside it isn’t a return. The same $11,000 is a gain against a $5,000 spend and a loss against a $20,000 one.
Before a vendor figure goes on a slide, ask four questions:
- What did it cost, including staff time?
- Whose organization, how many learners, and who paid for the study?
- Which lines are cash and which are hours?
- How was the platform’s share separated from everything else that changed?
A figure that can’t answer all four is marketing. It still has one use: the lines a vendor counts are a checklist of rows your own table might be missing.
Take the Table to Your Director, Not the Percentage
Build the table once the scoring matrix from your demos has a leader, and before you negotiate. Price your team’s hour at the loaded rate. Label every row as a quote, a receipt or an assumption, and split cash from capacity. Then write one sentence on where the freed hours go.
Report the payback month first, then each year of the contract on its own line. Turnover, productivity and errors go under the table, each with the method you’ll use to isolate it.
Do one thing first: time a real course build during the trial. Course conversion is the largest one-time cost, and it’s the one no vendor can fill in for you.
Mini Course Generator is our product. The only figure from it in this guide is its authoring-time claim, and it sits in a sensitivity line rather than the base case for that reason. Its plans and what each one includes are on the pricing page, so its subscription row doesn’t need an estimate.
Frequently Asked Questions
What is a good ROI for an LMS?
There’s no independent benchmark worth quoting, and the published figures mostly come from studies a vendor commissioned. A more useful test is whether the platform pays back inside the contract term you’ll sign, using lines your director accepts. In the worked example, that happens in month 16.
How long does it take for an LMS to pay for itself?
Two things set it: one-time costs and setup length. Count the running net month by month from the first invoice, not from go-live, because the setup months cost money and return nothing. In the worked example, $20,300 of one-time costs and a two-month setup put payback in month 16. Cut course conversion to the vendor’s claimed ceiling and it moves to month 11.
Should learner time count as a cost in LMS ROI?
Yes, when the training adds time that didn’t exist before. ROI Institute’s cost categories include participants’ salaries plus benefits for the time they spend in a program. When an LMS replaces training people already take, count only the change: a cost if the new course takes longer, a benefit if it’s shorter. In the worked example, an 8-hour classroom day that becomes 3 hours online is a 5-hour benefit for every new hire.
Is corporate training ROI the same as LMS ROI?
They share a formula but answer different questions. Corporate training ROI asks whether a program paid off. LMS ROI asks whether the platform decision pays off: what you’ll spend on it and what it replaces. If the question is still whether your company needs a platform at all, start with what an LMS does and when you can do without one.
Is a vendor’s LMS ROI calculator worth using?
Use it as a checklist, not an answer, and check what it leaves out. A calculator that asks only which use cases you have has no field for your own team’s hours. In the worked example, setup, course conversion and running the platform come to $25,520 of the $55,220 year-one cost, nearly half. Copy the benefit lines it counts into your table, then add those staff-time rows back.
Sources
- ROI Institute, overview of the Phillips model (five levels, BCR and ROI formulas)
- ROI Institute, The ROI Methodology in 12 Easy Steps, application guide (2019)
- U.S. Bureau of Labor Statistics, Employer Costs for Employee Compensation, June 2026
- Absorb ROI calculator (490% ROI claim from a Forrester TEI study Absorb commissioned), September 2026
- Docebo, How to Maximize Training ROI (MidFirst Bank example), September 2026
- Continu pricing page (implementation fee), September 2026
- Mini Course Generator AI Course Creator (instructional design time claim)
