Welcome to the TTA Community. TTA Connect is where you can manage and update your profile, search, and view opportunities, manage your work, track payments, and more.
The business impact of instructional design becomes easier to demonstrate when measurement is planned before development begins. This blog explains how to select a business metric, establish a baseline, track behavior change, and assign clear responsibility for collecting results. It also helps L&D teams decide when training is the right solution and when performance support may produce stronger, more measurable outcomes.
Decide how you’ll measure success before you build anything. During analysis, agree on the business metric, the baseline, the measurement window, and who will pull the data weeks before anyone opens an authoring tool. Wait until after launch, and you may be left reporting completion rates, a number that tells finance very little.
The evidence that instructional design can improve business results is stronger than many learning leaders assume. In a Journal of Applied Psychology meta-analysis of organizational training, Arthur, Bennett, Edens, and Bell found a sample-weighted mean effect of d = 0.62 for organizational results across 26 studies. The effect on job behavior was also 0.62, based on 122 studies and 15,627 participants. These are medium to large effects by commonly used standards in the field.
The evidence points to two conclusions. The second helps explain why learning teams keep having the same conversation about value.
First, well-designed training can change behavior and improve organizational outcomes by a meaningful amount. Second, relatively few studies collect that level of evidence. In the same meta-analysis, 234 studies measured learning, but only 26 measured results. When organizations stop at the knowledge check, they never gather the evidence that could settle the budget discussion.
That gap begins as a design problem, not a reporting problem. Kirkpatrick’s four levels have been part of the instructional design toolkit for decades, and TTA has written about evaluation as the fifth element of instructional design. Levels three and four are often skipped because no one planned for them while the program was being designed. By launch day, the baseline may already be gone.
Baselines cannot be collected retroactively. If you want to show that average handle time fell, somebody has to know what it was before the program ran, measured the same way, over a comparable period. Asking for that number three months after launch produces an argument about methodology rather than a result.
Analysis is also where you find out whether training is the right response at all. A proper needs analysis separates a knowledge gap from a process gap, a tooling gap, or an incentive that rewards the behavior everyone says they want stopped. Skip that step and you can measure the program perfectly and still watch the number sit exactly where it was.
One question can clarify most of this during analysis: If the program succeeds, what number will change, and who will notice? If no one in the room can answer, pause. The request may be a symptom of a problem that instructional design can’t solve on its own. Finding that out in week one protects the budget for work learning can influence and gives the sponsor time to redirect the request before doing so becomes expensive.
Let the business owner name it, not the learning team. It might be average handle time, safety incidents, error rates, first-year turnover, or quota attainment. Choose one so the program has a clear measure of success rather than an overloaded dashboard.
Capture the current figure before launch and agree how long you will watch it. Six weeks is too short for turnover and about right for a call-handling behavior.
Use a short manager-observation checklist, a sample of real work products, or a system report that shows whether people are following the new process. This evidence helps explain the business result, whether the number moves or not.
Decide who will pull the data and when the first readout will happen. A measurement task with several owners often ends up with none, so assign one person and one deadline.
For example, a contact center introducing de-escalation training might track repeat contact rate as the business metric, supervisor call reviews as the behavioral measure, and a short scenario assessment as the learning measure. Three months later, the team can see which measures moved and in what order.
Look for evidence that people are doing the work differently. A supervisor can review four calls against a five-item checklist in about twenty minutes, and the result is usually more useful than a 30-question survey about how the training felt. The same approach works in a back office: compare twenty escalation summaries written before the program with twenty written after it, then ask a second-level agent to score whether each one can be acted on without follow-up. Neither method requires a new platform, and both give the designer something concrete to evaluate.
When the work happens in a system, existing data may be the easiest source of evidence. Reports may already show whether required fields are completed correctly, approval steps are followed in order, or transactions need rework. Planning for that check is now part of content development for effective training, alongside the learning materials themselves.
Behavioral evidence matters because it helps explain the business result. If the metric improves, you can identify which behavior contributed and reinforce it in future programs. If the metric stays flat, the evidence can show whether people learned the material but ran into an on-the-job barrier. That shifts the follow-up conversation toward the manager, process, or tools instead of automatically blaming the course.
Choose performance support when employees need to look up information rather than memorize it. A 14-step process completed once a quarter, with no time pressure, is better suited to a job aid than a course. Turning it into a module may produce completion data, but little else.
Leiss, Rausch, and Seifried examined this question in Frontiers in Psychology through studies involving 301 HR employees and 652 ERP users. The availability of an electronic performance support system strongly predicted how often people used it. Support embedded in the work environment was especially well used when it was available. In practical terms, people are more likely to use guidance when it appears where they need it.
This choice can also simplify measurement. A decision tree embedded in an expense system comes with an obvious metric because coding error rates are already tracked and the trend can be reviewed a month later. When no one can name a measurable outcome for a course request, the real need may be support at the point of work.
Learning measurement covers the whole chain from reaction through behavior to business result. ROI converts the business result into a financial figure and compares it against program cost. Most learning teams get more value from a credible business metric than from an ROI calculation built on estimated dollar values.
Build one. Spend two weeks capturing the current figure before the program launches, or use a phased rollout so the groups trained later act as a comparison. A rough baseline gathered on purpose will survive scrutiny that a precise-looking number reconstructed afterward will not.
Rarely, in a corporate setting. A phased rollout gives you a natural comparison at no extra cost, and pairing a business metric with behavioral evidence is usually enough to satisfy a sponsor. Reserve formal control groups for programs large enough to justify the design effort.
Report the behavioral evidence. If people learned the material but didn’t change what they did on the job, the barrier may be the process, the tools, or management support. That finding gives the sponsor something useful to address, even if the program itself didn’t move the business metric.
Designing measurement into a program takes a different skill set from building the program. TTA matches organizations with instructional designers, learning experience designers, and evaluation specialists who set the metric during analysis and report against it after launch, drawing on a network of more than 5,000 instructional design professionals across 30 industries.
Find instructional design talent on TTA Connect
