Can you attribute business results directly to training?

Yes, it is possible to directly attribute changes in business results to specific training activities. What? Why are you shaking your head? Don’t give me that face! This is only the fifth sentence of the article. You can’t possibly be that skeptical already!

Man looking skeptical

Actually, I can’t blame you for your skepticism. Organizations have been struggling to measure the impact of training for years … decades … forever. The ROI of training remains trapped in a mythological land alongside the unicorn and sasquatch. We can all recite the four levels of the Kirkpatrick Model (reaction, learning, behavior, results), but we still can’t prove the impact of training on business results.

Meanwhile, across the office, every other department seems to have found a way to attribute their actions directly to changes in the business. Sales can do it. Marketing can do it. IT can do it. Why can’t Training do it? No, the answer is not “there are so many factors that influence performance.” Just as many (if not more) factors influence customer buying decisions, but Marketing can still attribute the impact of every ad they buy. What type of magic are they pulling off to make this happen?

Well, it’s not magic. The big difference between other functions and Training is that they work with A LOT more data. They can track small changes over time and create attribution models that determine the long-term impact of their actions. Marketing doesn’t use just one online ad to determine how they impacted you. They amass a significant collection of data on interactions and behaviors over time that ultimately connect to your buying decision. And that’s the problem with Training. They don’t have enough data!

More Data Required

To make attribution possible, Training must collect more data. But it’s not just about volume. They need to continuously collect the right types of data to tell their story. Right now, most Training departments collect the basics: attendance, test scores and surveys. This is the heart of the problem. You can’t tell if a course from six months ago improved sales results this quarter because everyone got a 100% on the test. That would be like trying to determine if a billboard on the highway impacted a consumer’s decision to buy a car 8 months later. This just isn’t enough data!

The 5 Vs of Big Data

Training must expand the definition of “learning data” to include an array of metrics that measure the full spectrum of performance changes over time. These metrics include:

  • Consumption: what training resources employees are using
  • Knowledge: changes in what employees know over time
  • Confidence: how ready employees feel in their ability to apply their knowledge on the job
  • Behavior: changes in employee actions on the job over time
  • Results: key performance indicators for the business that are targeted by training programs

A New Data-Rich Experience

So how do you collect this type of data? If you want to create attribution, you have to change the experience. Think about the changes in digital marketing over the past 20 years. The increase in data was the result of new forms of advertising, especially online. These innovations created new opportunities for data collection and resulted in a much stronger attribution model.

Similarly, Training must evolve the learning experience for employees. Not only does this provide additional data collection opportunities, but it also aligns to the way people really learn – in small amounts over time. By introducing the habit of everyday learning through strategies like microlearning, organizations can measure what employees know and how that knowledge is being used on the job. Rather than make guesses with limited data, like test scores and completions, Training can analyze more targeted data points that show employee knowledge and behavior changing incrementally over time alongside related business metrics. That’s how you create a learning attribution model.

The continuous learning experience
Continuous learning experience


Building Training with Measurement in Mind

Training content is another important consideration when it comes to establishing attribution. The more targeted the content, the better the data. Rather than building bulky courses on a wide range of topics, Training must use a results-first approach and deliver content on only topics that will impact the selected business outcome. For example, rather than create a course on “Safety in the Workplace,” Training should identify the most critical safety issues in the organization and build targeted solutions. If back injuries are a major problem, they can build a reinforcement module on lifting procedures rather than bury this content in a longer course. In doing so, Training must identify the intended result (reduce back injuries), required behaviors (proper lifting techniques) and fundamental knowledge (lifting dos and don’ts). Each of these components can then be measured after the training is delivered to determine changes in knowledge, behavior and, ultimately, results.

Axonify's results-first approach
Results-first approach to training


Partnership is Critical

Training should not attempt to make this data transition on their own. After all, the smartest data people in the company are probably not in the Training department. Rather, Training should partner with other teams who have experience with advanced analytics and access to additional types of data. This includes Operations, Safety, Sales and Business Intelligence. By bringing together data from these various sources, Training can create stronger attribution models that connect their efforts to business results.

Going Beyond Attribution

Determining the impact of training on business results is just the first step in the process. Once Training knows what is and is not working, they can make proactive decisions to improve their solutions. This data can also be used to personalize and adapt the training experience for the individual employee so they stay focused on topics that are proven to have the greatest impact. Frontline managers can use this data to improve their coaching and help their teams build the knowledge and skills that are proven to improve business results. At Axonify, we’re taking the idea of learning attribution a step further by using machine learning to help organizations improve the impact of their training on business results.


So the answer is a clear and definitive YES: you can attribute changes in business results directly to the impact of training. And if you still don’t believe me, check out this case study from Bloomingdale’s to see how they used the strategy I just explained to attribute $2.2 million in savings per year to their microlearning strategy. It’s time for Training to catch up with the other departments in harnessing the power of data so they can focus on solutions that are proven to have real impact on the bottom line.

JD Dillon became an expert on frontline training and enablement over two decades working in operations and talent development with dynamic organizations, including Disney, Kaplan and AMC. A respected author and speaker in the workplace learning community, JD also continues to apply his passion for helping frontline employees around the world do their best work every day in his role as Axonify's Chief Learning Architect.

Let’s work together to drive frontline performance in all the right ways.