The space-race to build the Netflix of Learning

Racheal Reeves
17 min readJul 31, 2019


How did I end up attempting to build the Netflix of Learning?
To start, I am not your typical human. I once worked with a psychologist who suggested that people with brain injuries have more inhibitions than I do. And secondly, I am not a Learning and Development specialist. I have a background in technology and design.

I discovered Learning and Development four years ago. It was amazing. I started to read, watch and listen to understand how people learn. I connected all this new information to things I have previously experienced or learnt. And finally, I concluded that in every decade of contemporary history, we had applied technology as the solution to the same problem.
How do we get people to learn?

Today, the newest and latest technology is Netflix. But does it solve the core problem and what is the core problem? I see the core problem as:

How do you change a persons’ state from employee to user to learner to high performer?

This story starts this time two years ago, when several Learning & Development teams across Spark became one big family. Between them, we had several learning platforms, and we needed to reduce complexity and optimise the experience — ideally with a system as cool as Netflix. As we were looking at what to do about our multiple LMS’s, we came up with a pretty long shopping list:

We shopped around, and many people thought we were quite insane with our long list of wants and our small pocket of coins. In the end, we could not find a single LMS that would allow us to drive learning suggestions to people using their performance data. We decided to build our own using a Content Management System rather than a Learning Management System.

There are a few key reasons we did this. The biggest one is because of supply and demand. Moodle is an extremely popular LMS has 105,000 registered sites worldwide. Drupal is an extremely popular CMS has 965,000 registered sites worldwide. Both platforms are open-source, which means developers of these platforms share their code.

Open source gives you the ability to build platforms like a jigsaw puzzle without expensive customisations. Now would you rather have a platform that was supported by 965,000 other platforms or supported by 105,000? Our very tight budget knew which one it preferred. So the space-race to build the Netflix of Learning began.

What is Netflix, and how does it work?
It is easy to explain a world before and a world after Netflix.

The world before Netflix was about working for your entertainment. It involved TV channel guides, recording shows, watching adverts and arranging your life around the TV shows you wanted to watch. If you wanted to see a movie, then you got in your car and drove to the local Video store. The quality of your movie-watching experience was not constant — if you hated the video, you couldn’t switch to another. If the film was a videotape, then you may need to rewind it or if it is a DVD then it might have scratches on it. All these scenarios would reduce the quality of the movie-watching experience.

Post Netflix you can find a movie or tv show that perfectly matches your preferences and all for the cost of one “recent release” video rental.

The success that Netflix has enjoyed has meant that other industries are looking at how they can mimic their business model. Everyone wants to harness the awesome power of Netflix. So, naturally, we want it as well. But what is the Netflix of learning? And how do we apply the Netflix methodology to learning?

There are a few key things we need to look at to realise our dream of harnessing the power of Netflix in our organisations for learning.

Firstly, let’s look at the critical difference between Netflix and Learning platforms, the content. Netflix has entertainment content, which is passive. This type of content requires the user to do only one thing — consume it. Learning platforms have active content. The user may need to process, store, apply, contextualise and blending the content with other material, the list goes on.

Have you ever tried to watch a Ted talk, listen to an audiobook or started watching an instructional video when you haven’t had your brain switched on? It’s like being in a strong current, and you are getting washed downstream with no control. You are overwhelmed by the water gushing over and around you that when you get to the quiet, still lake at the end, you have no idea what just happened. You are just glad you made it out alive. You are also now very apprehensive about diving into that river again. That’s what it’s like if you try to consume active content in a passive state.

If you approached that same Ted talk or instructional video with an active state, it’s much like going, okay I am ready to tackle this river. I have had my breakfast, I know where to start in the river and where I want to end up, I have my kayak, my paddle. I am ready. Then as you are going down the river, you are analysing that river, taking the learnings from each stroke of your paddle to make a better stroke the next time. But it’s not often you get out of bed and go “yes, let’s go kayaking.” You have thought about going kayaking, made time to go kayaking, previously acquired the things you need to go kayaking.

Let’s run a test to see if this is true:
* Did you go kayaking last weekend?
* Did you watch Netflix last weekend?

How could I get you all to kayak every weekend? What would convince you? If we used the Netflix model on kayaking what would happen? The system would gather all kayaking content (equipment) and connect it to a user based on their preferences. All the user would need to do is push a button, and it would be at the river ready for them. But would this encourage you to get up every Saturday at 6 am and drive to the river?

No, not unless you want to kayak in the beginning. It may persuade those who like kayaking. People will not be convinced to step out of there comfort zone and invest time and energy in kayaking by a frictionless experience to collect kayaking gear. The magic of Netflix can only go so far, and we need to think about more than just using Netflix.

What currently stops people from entering a learning platform is much the same reason we don’t want to get up at 6 am to go kayaking. The perceived costs of using your platform are higher than the value they get from it. In other words, the time, energy, mental real estate that a person spends far exceeds what they get out of the experience of using your learning platform.

What do I mean by value?
At work, it could be things like:
* Get what they need to get done with the hours they have
* Advance in their career either by pay, recognition, respect or title.
* Solve problems or difficult challenges
* Work out ways to reduce the components of their job they do not enjoy

For example, imagine if you hated the administration that went along with project management. You may be motivated to learn a new management tool that will reduce the amount of administration you need to do. You will learn how to use this system if you see the value of the outcome exceeds the time and energy you put into learning it. So how can we change people’s minds and show that learning platforms are valuable?

In UX design, there is a model called the hook loop. It is based off how the brain creates habits and is a time-tested way of getting users hooked on a game or product. It’s an incredibly simple concept with proven results. It is made up of four things, trigger, action, reward, investment.

This diagram was taken from which is a great article so go read it (after this one of course!)

When we designed our new learning platform, Ako, we put hook loops at the core of the experience.

It starts with a trigger that prompts the person to enter the platform. We looked at what would motivate our people to want to join the platform. What would make it of value to them? Instead of using a marketing campaign, we drilled down to the heart of what our people do day-to-day. They work, perform and face challenges. With this in mind, we exposed our peoples’ performance on the platform.

Once we got our people into the platform, we needed an activity for them to do. Much like how you get people to do an energiser activity at the start of a workshop to get peoples brains flowing. Its something low effort but breaks the ice. Our action was to click on your performance dashboard.

When you open the dashboard, you are immediately rewarded with your performance metrics and clear pathways on how you can improve. It gives you a hit of dopamine or adrenalin depending on your expectation’s verses reality. Both situations break the ice.

You are now in a position where you are willing to invest your time and mental energy into performing better. The great thing about a hook loop is that each time a person goes through the loop, the time and energy they invest gets exponentially more significant.

The investment is the associated learning to increase your performance to get to the next salary bracket.

Each month you will return to view your updated dashboards and see, for example, you qualify for the next salary bracket establishing the value of learning. Learning equals performance and performance equals money.
For learning, we want to develop a double hook loop as we want our people to invest in both learning and in applying the learning to increase performance.

So let’s go back to the ultimate problem we are trying to solve:

First off, we have understood our people and triggered to get them into the platform where they have become a user. Now we want to change their passive user state to an active learner.

Its the big jump between enjoying Netflix and going kayaking.

Ako was built in five weeks and went live in April last year. Since then, we have done hundreds of deployments to enhance, fix and create functionality that driven by our peoples’ feedback. Each iteration has allowed us to collect data and understandings of what our people want, and how they use the system. It also showed us that building the Netflix of Learning doesn’t happen overnight. It takes time, data, commitment and careful resource allocation.

In each iteration of Ako, we had a minimum of one problem we knew we were going to solve and one experiment. We were working on tight budgets with a team already at full capacity. The first thing we needed to do was remove a whole lot of work off our plates so that we could free up time to innovate.

We analysed the core frictions in our previous system and isolated these to be:
* The system required a steep learning curve to use
* Content uploading was labour intensive and complicated
* Report pulling was strenuous and could only be done by a few
* Administrators needed to register users on courses

It was clear to us that our first experiment needed to benefit us. We needed to make the administrative side of the system simple to use. We developed a user permission series that works as a network. It was a roaring success. We went from pulling thirty-odd reports a week to not pulling any.

The second core component of the first iteration was surfacing all content to everyone. Spark has multiple different customer channels and different areas that required different content. We decided to create a transparent platform where anyone could access any content. We categorise content by connecting it to relevant KPI’s to create a direct connection between learning and performance.

The performance dashboard in the second iteration created a space where our people have the autonomy to analyse their data and understand how o succeed in a safe space. This safe space engages and encourages you to set your performance goals.

We finally achieved to shift our people from a passive state to an active state. And how do we know this?

We surpassed the engagement rates we had in our previous platform. We saw a higher return rate, an increase in the length of stay and an increase in the amount of content consumed. We also drove an increase in performance — in some KPI’s we saw a 500% increase and contributed to an overall rise in customer experience for New Zealanders.

When we first connected our content to the KPI’s to help people learn, we saw a massive increase in the consumption of these modules. But then our metrics started to show a decrease in elearning consumption for performance-based elearnings. Instead, people were participating in side-by-side activities or conscious practice at work to develop their skills. Now we had to ask the question — is our content the right content?

Now, what would Netflix do?

One of the smartest things Netflix did was commission content based on analytics to ensure they were giving their viewers what they wanted to see.
Before paying 100 million dollars for two seasons of House of Cards, Netflix wanted to be sure that the show was a sound investment. Instead of watching it and deciding if they thought it was good, they looked at the data they had collected on their subscribers. Inside the House of Cards package was three key data points:
* Netflix knew the British version of House of Cards was well watched
* David Fincher is the director
* Kevin Spacey was the main actor
Alone, these points are not very exciting. When Netflix combined the data sets, they found that people who watched the British version of House of Cards also watched Kevin Spacey films or films directed by David Fincher. House of cards became a roaring success.

We can use this same concept to make our learning content sing. We can work out what people want, how they learn, and what content is the most effective. But before we could do any of this, we needed to create a scenario where we could capture all of this data. To create the Netflix of Learning, you need lots of data points.

The first thing we did was audit all of our content. I am forever grateful for the help we got on making this audit happen. So thank you again, everyone, in Learning Enablement. We all got stuck in, reviewing content and tagging it with as many data points as possible. Data points can be used for a multitude of things. And expand your analytics game from the minor leagues to the majors. They are like hashtags and allow content to flow through different information structures.

Three key reasons to populate data points on the courses:
* To recommend personalised content to a user.
* Understand what content works and what content doesn’t work to drive better content creation.
* To create a user-friendly system where people can easily find what they want.

You can do this by tagging your content to lots of different things, like the author, date it was published, the time it takes to complete, performance metrics, people metrics, content types. The more things you can think to tag it with the more power you will have over the content. Data points are how you connect users to content to create a Netflix of Learning.

We also wanted to see how people moved around our platform. So we plugged in google analytics, which is extremely easy to do on an open-source piece of software. It’s free, and it can tell you everything from the behaviour of a person on your platform to what page they exited on. Our analytics said modules are too long.

We had 40min modules, but our people spent approximately 2 mins in these modules. We did an A/B test with a new series of microlearning modules which were approx 5 mins each. We discovered that the average amount of time people spent in the micro modules was five minutes. This data helped us reduced the amount of time we spent developing modules and freed us up to do even more cool stuff. We went from the launch of Ako having users on the platform for approximately seven mins to now having an average of ten minutes within the space of a year.

The second change we made is to the way we create content. We reduced content creation by Instructional designers and instead created a network of peer-to-peer knowledge sharing. Our designers switched from creating single modules to supporting and developing our people to build modules.
We again checked with the data to see if this was working. We removed mandatory modules outside of compliance and found that the modules with the highest level of engagement through time and completions are the ones created by our people for our people.

For instance, I was designing an elearning module that would help our frontline with Fibre orders. I had no idea how a conversation about Fibre. So, I reached out, asking one of our frontline teams if they could help me. The team jumped into it and created epic videos themselves. That module ended up having a considerable amount of positive feedback — better than ones solely created by instructional designers.

By using the insights that data provide you — you can change the way content is developed and ensure that your people become users who then turn into learners.

But what happens when our people don’t do what we think they will do?
A few months ago, I didn’t feel like working. I felt drained, over it and classically brain dead. So I went for a swim. I do this a lot. It’s my escape from technology as you can’t bring it into the water with you!
And there I was doing laps in the slow lane, and my brain switched over to nothingness and then it got exciting. As I zoned into swimming, my mind started firing through how to solve a problem I had been stumped on. It was while I was swimming lengths I pulled my learning together to apply it to my job.

We cannot demand that learning can only be consumed by our teaching, or by the learning outcomes we define. If we remove autonomy from learning, we remove the ability to learn. We remove the motivation to explore, and we create learned helplessness.

But how do we ensure that we are doing our jobs, increasing performance — if we let people swim around? It’s a pretty terrifying concept. How do you know I wasn’t thinking about kittens while I was swimming? And how do you know if I applied my ideas to work? And the answer?

The proof is in the pudding. Delicious pudding.

We created a program called Accreditation a career framework based on performance metrics entirely driven by hard data that comes directly from our data warehouse. This created a clear, non-biased base-line that our people can measure their success off. It is clear, transparent and tracks all increases and decreases in your performance both holistically and at a single KPI level.

Accreditation shows you how well you are performing in each of your KPI’s and what overall level is so that you can see your progress.

Accreditation became a benchmark for if a person is performing well instead of completion results. It also meant we could give our people real feedback about the development they are doing and allows them to create goals out of this data. It provided you with the freedom to work out how you are going to grow, develop, learn and perform. If swimming is what works for you — go for it.

We have for such a long time struggled to correlate data with learning, but Accreditation has bridged that divide. It has given us the data to prove that we are right. It has told us to give people autonomy to learn the way they want to learn and to communicate performance metrics.

Our next steps are to integrate this into our Netflix of learning to continue to improve our content. We already link learning suggestions to each KPI and are at the point of almost having enough data to start acting this.

It’s important to remember that Netflix gathered billions and billions of data points to come up with conclusions. So, don’t be impatient. Build fast and build cheap to collect your data. And then slow down.

We did a 5-week build that stripped us back to the bare bones of an LMS, but we increased the ability of what data we could collect. Every function we have built has amplified our impact and has also told us when we have screwed up.
We could never have achieved the success we have if we had not built in an iterative approach, focusing on collecting data first. Through constant feedback, we understand what our people want. For instance, we know they want to search the entire site, not just course titles, which means we need to do two things:
* Phase-out SCORMs as they are unsearchable
* Write transcripts for videos, so the words are searchable
* Which is our next mission!

When we look back at the core problem, how do we get people to become a user to become learners to become high performers, it shows up like this:

Employees are people
Focus on what’s important to them — how they can solve their problems and challenges and progress in their careers.

Reward their curiosity of seeing their dashboards with things they can do to increase their performance

Create learners by empowering people to develop and achieve their own goals through the transparency of data

High performers
Enjoy a better workforce that is more engaged, focused and autonomous by giving them the tools and space to create their meaning.

Here are the rules that we have lived by to create Ako, our people’s platform:
1. Create platforms that create autonomy, and then you won’t need to force people to hate their way through training.
2. Create platforms that celebrate successes and help identify problems in a safe space.
3. Create platforms that feedback what a person does matter — help a customer with a particularly trying problem. See your FCR results increase as your customers respond to your hard work.
4. Create systems that are created by your people. It’s their lives, their content, their dashboards — not yours.

There are many different ways you could tackle the problem of how to increase performance in your organisation with technology. I would love to hear what you think about my approach and what you are currently doing. Get in touch!