Sales training ROI: how to measure it on calls

Measure sales training ROI with the formula, the four levels and behavior on calls as the early signal. A worked example and what software can measure.

Hank Mesquita

11 min read

Sales training ROI is the money value of what the training changed, minus what it cost, divided by what it cost. The formula is simple; the work is in measuring the change and proving the training caused it, and the earliest place you can see that change is on your team's calls, weeks before it reaches revenue.

This post covers the formula, the levels of training evaluation, why behavior on calls is the leading indicator, which calls to measure (with our own pipeline data by call type), a worked example and what software can measure for you. For a general six-step method that applies to any enablement program, read how to measure sales enablement success. This post narrows it to training.

What is the formula for sales training ROI?

The ROI Institute, which publishes the ROI Methodology, writes it as ROI (%) = net program benefits / program costs, multiplied by 100. Net benefits are the money value of the gain minus the costs. The same page lists a second ratio, the benefit-cost ratio: program benefits / program costs.

The costs are easier to undercount than the benefits. Count all of them:

CostHow to estimate it
Trainer or contentThe invoice, or the hours your team spent building it
SoftwareSeats for the period you measure
Selling timeHours in training x people x a loaded hourly cost
Manager timeHours of review and coaching tied to the training
Travel and eventsOnly if the training was in person

Selling time is the line buyers of training notice most. A sales operations leader told us on a call in July 2025: "we're also balancing always is the cost of the training, time they're not making potential sales." Put it in the cost line, and the ROI you report will survive the CFO's first question.

Which levels of evaluation should you use?

Use four levels, and add ROI as a fifth when someone asks for a money number. The four levels come from the Kirkpatrick Model, which its site says Dr. Donald Kirkpatrick created in the 1950s: reaction, learning, behavior and results. The ROI Institute's methodology uses five levels, with return on investment as the fifth.

LevelThe questionSales exampleWhen you can see it
1. ReactionDid people find it relevant?Survey after the sessionThe same day
2. LearningDo they know it?Quiz, certification, a scored role playThe same week
3. BehaviorDo they do it on the job?The step is met on real discovery callsTwo to six weeks
4. ResultsDid the business number move?Discovery-to-demo rate, win rate, ramp timeOne sales cycle or more
5. ROIWas it worth the money?Net benefit / costAfter level 4

Many training reports stop at level 2, because completions and quiz scores come out of the training tool for free. Level 3 takes more work, since someone has to check real calls, and it is the level the rest of this post is about.

Why is behavior on calls the leading indicator?

Because it moves first, and every later number depends on it. A win rate needs a full sales cycle to move and is pushed around by pricing, territory and the market. The steps on a call change within weeks and depend mostly on the seller. If the step didn't change, the training can't take credit for any revenue that follows.

Leaders describe the missing level 3 on calls with us in plain words. An enablement leader said in May 2025: "it's so hard to get the read of whether or not it's landing." Another enablement leader described what happens without it, in March 2025: "we train multiple times on the same topics. At least once a quarter we're having to retrain." These are teams already looking at coaching tools, so read the quotes as examples, not a market survey.

Jacco van der Kooij of Winning by Design explains why the clock matters in our session on metrics of productivity. Citing an average tenure of 12 to 18 months in software sales, he says:

"These metrics tell me that I need to get people ... productive in three to six months, otherwise there is not a lot of value that they have. There's not the time left for them to provide the value."

He also names what to teach to get there: "teach the use case stories, teach them the impact that your solution provides and a critical event that you're selling against." Each of those can be heard on a call, which makes each one measurable at level 3.

The link from level 3 to level 4 can be tested with your own data. We did it on Triple Session's own sales pipeline in HubSpot: 208 deals marked closed (61 won, 147 lost), with close dates from July 2025, read on October 28, 2026. We took the first scored sales meeting of each deal and compared its score with the outcome.

First meeting score and deal outcome in our own pipeline
First meeting scoreDealsWon
Under 807428%
80 or more6343%

Sample: Triple Session's own sales pipeline in HubSpot: 208 deals marked closed (61 won, 147 lost), with close dates from July 2025, read on October 28, 2026. After cleaning, the first scored sales meeting of each of 137 deals. Method: First meeting scored against our own scorecard in Triple Session, matched to the deal outcome in HubSpot. First meetings only, to remove the bias that lost deals die after one meeting.

The score and the outcome move together, and that is all this shows. A good lead also makes a better meeting, so the score predicts the outcome, and raising the score does not by itself make a deal close. It is one company's pipeline, our own. What it shows is that a call score can work as an early signal for results, which is the bridge a training ROI needs. Run the same comparison on your own deals before you rely on it.

Which calls should you measure for level 3?

Measure the calls where the score tracks the outcome, and skip the ones where it doesn't. Not every call type works as an early signal. In the same pipeline, we split the scored meetings by call type:

Call typeMeetingsScore under 80: in a won dealScore 80 or more: in a won deal
Discovery14323%40%
Demo12037%54%
Follow-upn/a71%69%

Discovery and demo scores separate won from lost deals. Follow-up scores don't, because a deal that reaches a follow-up is already close to the end. For a training ROI, that means the level 3 measure belongs on discovery and demo calls. A training on follow-up calls can still be worth running, but its step rate will not tell you early whether revenue will follow. Same caveat as above: correlation, one company's pipeline, and your call types may behave differently, so run the split on your own deals.

The worst point in our data is also the easiest place to show a return: discovery meetings scored under 60 became a sale 4 times out of 39 (10%). A training aimed at the steps those calls miss has the most room to move.

What does a sales training ROI calculation look like?

Here is a worked example at Larkspur Analytics, a fictional software company with 12 account executives. The numbers are invented to show the method. The team runs a discovery training on one step: putting a number on the cost of the buyer's problem.

Step 1: cost.

Cost itemAmount
Outside trainer, one day plus materials$18,000
Selling time: 12 people x 6 hours x $150 an hour$10,800
Manager review: 2 managers x 2 hours x 8 weeks x $100$3,200
Total$32,000

Step 2: behavior (level 3). Six people get the training in January, the other six in April. Before the training, the step is met on 30% of discovery calls in both groups. Six weeks later it is met on 62% of the trained group's calls and 33% of the others'.

Step 3: result (level 4). The trained group's discovery-to-demo rate goes from 40% to 48%. The untrained group's goes from 41% to 42%. The training gets credit for the difference: 7 points.

Step 4: money. After both groups are trained, the team runs 180 discovery calls a quarter. 7% more of them become demos: 12.6 demos. At a 25% demo-to-close rate and a $24,000 average deal, that is about 3 extra deals, or $75,600 in a quarter.

Step 5: ROI. ($75,600 minus $32,000) / $32,000 x 100 = 136% in the first full quarter. The benefit-cost ratio is 2.4.

Two notes on the example. The control group is what makes the 7 points believable; without it, the team would credit the training with all 8 points, plus whatever the market did. And if finance counts margin rather than revenue, multiply the $75,600 by your gross margin before you divide.

How is this different from the usual ROI advice?

The common advice on this topic either lists ways to raise the ROI or shows a before and after calculation. LSA Global lists seven strategies, starting with aligning the training with the sales strategy. Menemsha Group works through sample scenarios for staffing firms, one of them reaching a 3,700% ROI.

Both are useful starting points. The method above adds three things: a behavior measured on calls between the training and the result, a choice of which calls to measure based on where the score tracks the outcome, and a group that hasn't been trained yet. The first tells you early whether the training is working. The second keeps you from measuring calls that can't move the number. The third tells you how much of the result to credit.

What can software measure for you?

Each kind of training tool measures a different level, and none measures all five alone.

ToolLevels it measuresWhat it gives you
Learning platform or LMS1 and 2Completions, quiz scores, certifications
Practice and role play2Scores on practice calls
Call scoring3Steps met on real calls, per person, by week
CRM4Conversion, win rate, deal size, ramp

The ROI calculation needs level 3 and level 4 in the same spreadsheet. If your call scores and your CRM can't be joined by person and date, you will end up reporting completions. Our post on sales training tools groups tools by the revenue number they move, and sales enablement KPIs covers which metrics to track at each level.

I work at Triple Session, which scores real calls and practice calls against your steps, which is the level 3 data in the table above.

Which mistakes make the ROI number wrong?

  • Reporting completions as impact. A finished course is level 2. It says nothing about calls or revenue.
  • No baseline. Without the step rate and the conversion rate before training, there is nothing to compare with.
  • Crediting all of the change. Pricing, new leads and a good quarter move numbers too. A group trained later is the simplest control.
  • Measuring too soon. Revenue needs a full sales cycle. Report the behavior at six weeks and the result after the cycle.
  • Leaving out selling time. It is often the biggest cost, and the one a CFO checks first.

Frequently asked questions

How do you determine ROI for training?

Subtract the cost of the training from the money value of the gain it caused, divide by the cost and multiply by 100. The hard part is the gain: measure the metric before and after, compare with a group that did not get the training yet, and credit the training only with the difference.

What are the key goals of sales training?

To change what people do with buyers so that a business number moves: more first meetings that turn into opportunities, a higher win rate, larger deals or a faster ramp for new hires. A goal is only useful if you can name the behavior on calls that should change and the number it should move.

How do I calculate the return on investment (ROI) for sales?

ROI (%) = (gain from the program minus its cost) / cost x 100. For sales, turn the gain into money with the metric it moved: extra deals won x average deal size for win rate, or weeks saved x weekly output of a ramped seller for ramp time. If finance counts margin, apply your gross margin to the gain.

Where to start

Before your next training, write down one step it should change and one number that step should move. Measure both now, train half the team first and check the step on calls after six weeks. If it moved, wait a sales cycle for the number. For the measurement plan around it, see how to measure sales enablement success, and for how training and coaching work together so the step sticks, read sales training coaching. The online sales training platforms guide covers what to look for in the tools.

Sources

All read on November 16, 2026.

See your team's calls scored on your playbook.

Send a few recordings. We score them and show where each person can improve. No cost, no contract.