Why so many digital ELT businesses fail to make money, and what we can do about it

A recent survey of failed startups found that ‘business model not viable’ was the single most common reason for failure. To date, a product’s business model has been the preserve of the small group of executives or founders who own the ‘business strategy’. Monetisation isn’t something that the people actually developing the product typically concern themselves with. As a result, product development is very operationally focussed. Lean is a process. Agile is a process. They are formulas for developing software products. But these processes, blindly adhered to, have inadvertently created a glut of digital products with business models similar to the South Park gnomes’ infamous Profit Plan. Compare:

Collect underpants → ? → Profit


Build product → Pivot ?— -> Profit

paul 1

With oversupply the norm in almost every conceivable product category (the ELT content and services market included), it’s not building great products that’s our biggest challenge; it’s monetising them. If there ever really was an era in which you could pivot to profit, it’s over. Instead, those of us developing digital products could do worse than to follow the advice of Michael Porter, the Godfather of Strategy, and concentrate on identifying what makes their product unique and therefore defensible. In this post, I explore why monetisation is such a chasm to cross and the decisions I believe we should be making upfront before building anything at all. In doing so, I’ll be drawing on the ideas of the current ‘capo di tutti capi’ of product strategy, Ben Thompson. Ben’s epic project, Stratechery, demystifies and deconstructs the digital product landscape and the hidden forces at work within it. If you’re interested in why some startups evolve into unicorns while others sink into obscurity, you’d be well advised to familiarise yourself with Thompson and his work (a unicorn is a company with a valuation of over $1 billion).

As the name indicates, Stratechery focusses on the big strategy of big tech. Like all great theorists, Ben stands on the shoulders of giants. He absorbs strategic thinkers of the past  –  like Michael Porter and Clay Christensen  –  and identifies why they are still relevant and where the new reality of digital exposes shortcomings in their thinking. Despite an explosion of product and technical innovation during the last decade, successful models for monetisation have remained surprisingly static. There are only a handful of ways to monetise digital products, and some are more equal than others. Whilst some categories (like games) have been able to develop micro-payment models (via in-app purchases), for the majority there are only two business models that have proved to be repeatable: advertising and subscriptions (Software as a Service, or SaaS). Note that selling content is not among them.

Deciding which model is right for your product depends, in turn, on another critical choice: whether your product is a consumer (B2C) or a business (B2B) product. In ELT, this might be simplified to: are you selling directly to students or are you selling to institutions?  This choice is fundamental to product strategy because your business/monetisation model informs every aspect of how your product needs to be designed and built. Institutional buyers have very different needs from students, and it’s easy to waste precious resources building for the wrong market. As the survey showed, most startups die building a product for a user that will never pay.

Consumer products and the (big) problem of free

‘People have no problem in paying $5 daily for a coffee. But they have a problem in paying upwards of $0 daily to use an app. Heck, nobody’s going to pay to use your app for $5 a year, let alone a single day.’

Mike Post  –  FitFriend

In his 2014 piece on business models, Ben Thompson concluded that ‘the removal of friction from the software buying process has reduced or removed the willingness of many consumers to pay’.

Why is this the case? Because consumers have an innate sense of marginal cost – i.e. how much it costs producers to replicate their product for each new user. In software, marginal cost is $0. Consumers know this, so are unwilling to pay. An entire generation of digital consumers has grown up listening to icons like Mark Zuckerberg and Sergey Brin proclaiming that software should be free.

Graph showing that for digital products, price tends towards zero.
With thanks to Ben Thompson.

It follows, therefore, that the most robust way to monetise is by advertising to a user base paying $0 for your product. Let’s face it, advertising is a big deal in digital – it’s the rocket fuel of unicorns. For category-defining, must-have products like Snapchat, YouTube or Facebook, consumers have consistently demonstrated they are happy to tolerate advertising in exchange for getting these experiences for free. This is not a model completely alien to ELT; for example, we’ve seen some publishers have a level of success with generating advertising revenue from their dictionary websites. But advertising is a complex beast. It’s not as simple as ‘build product, scale … then add advertising’; that’s (another) Gnome Profit Plan. Consider this: in July 2016, Twitter reported 313 million Monthly Active Users (MAUs). Yet it is in deep, deep trouble – a basket case of missed earnings and failed expectations.

Twitter is not alone. Many products (or companies) with Active Users in the 00s of millions are still struggling to break even. Think Busuu (60 million registered), Duolingo (150 million registered) or even LinkedIn (433 million!). All have audiences most can only dream of … yet they still struggle to make their business models work. The reason why can be explained by another stat from Mary Meeker’s annual Internet Trends update in June. In 2015, Google and Facebook combined captured 76% of internet advertising spend. They achieved this by capturing an equally disproportionate amount of the mobile app market. In fact, in 2015, Google and Facebook claimed the top 8 mobile apps!

Advertising Revenue and Growth Rates (%) of Google vs. Facebook vs. Other, USA 2014-2015
With thanks to Ben Thompson.

These eye-popping stats are terrifying for ‘old’ media publishers (ELT publishers included) and equally foreboding for just about anyone looking to monetise a product via advertising. Confirming Meeker, the overarching message from 2016’s Cannes Lions Festival of Creativity was that we are now in the age of an advertising duopoly … and their grip is tightening.

If you want to monetise through advertising you need to focus on building a specific type of product. For a start, your product needs to be sticky. That means ‘several times a day’ sticky, at least. Don’t forget, over 80% of people’s time with their phones is spent in just three apps. You need to have a realistic chance of claiming one of those spots or you’ll never get the necessary traction.

Nir Eyal’s habit-forming, hooked model is your playbook here. You need to understand the psychology of habit and addiction. Then you’ll need to deploy every notification, web hook, update and FOMO-inducing trigger imaginable to keep users engaged and coming back into your product. The seemingly unstoppable demolition of Twitter by Facebook is still the best case study on what happens if you get this right.

Next, you’ll need deep data insight into your users and their demographics. Advertisers pay for access to very specific audiences. You need to capture that data somewhere in your onboarding and be able to demonstrate quantifiably who is using your product. You’ll also need to build out ad-serving capabilities, conversion tracking and reporting sooner rather than later.

Finally, you’ll need to ensure your product is set up for instream, immersive advertising, not ugly interstitials (adverts that appear while a web page is loading). This is why Snapchat is winning  –  it’s perfectly constructed for the full-screen, video placements beloved of brand advertisers.

To have a hope of attracting even a small percentage of the ad dollars that the ‘Foogle’ duopoloy leaves on the table, you need to consider how this will work from the outset – not as a bolt-on or as an afterthought.


Enterprise and SaaS

Ad-supported products are, by definition, consumer-focussed. This is primarily because ‘no business is going to rely on an advertising-based service for critical line-of-business needs. Beyond the lack of professionalism, data security concerns makes any consumer’s service a non-starter’ (Ben Thompson).

Critically, enterprises are NOT starting from an expectation that software should be free (although whether this is true for small-scale institutional customers in ELT – such as private language schools – is a matter of debate). Consumer SaaS (for example, selling subscriptions directly to learners or teachers) is still a tough gig, as many in ELT have learned. Slack or Box are more realistic role models than Netflix or Spotify, so B2B is a better bet for those people developing digital products looking to monetise via subscription payments. One of the biggest problems in consumer SaaS is how customers pay. PayPal and credit cards are still rare in many regions and local payment options can be complex. In the enterprise, circumstances are reversed – not only are customers set up to pay; they (mostly) expect to do so.

Product development for the enterprise has specific characteristics and considerations which, for people used to developing products for consumers, can be counter-intuitive. First up is the significantly reduced premium placed on the user experience. Switching costs are more pronounced in the B2B environment, so design doesn’t factor into a purchase decision. A detailed cost/benefit analysis, however, most definitely does. Lest we forget, the buyer is rarely the user in the enterprise, so the purchase is a rational calculation, NEVER an emotional one. This is why Apple hasn’t penetrated the enterprise the way Microsoft has. Might this also explain the existence of so many badly designed LMSs? The people responsible for buying them knew they would never actually have to use them.

As anyone who’s ever sold Business English courses into enterprises knows, business buyers are looking for measurable outcomes that demonstrate a return on their investment, so your product must have progress tracking and detailed reporting metrics from the outset. Businesses also have specific customisation or feature requests, and are willing to pay for them. So those developing digital products must be willing to weigh the business case for custom development against competing demands from other stakeholders.



Michael Porter says strategy is about making hard choices. In this post, I have made the case that people developing digital products need to make hard choices about their business model before building a product that will sustain it. For most of us, that means an either/or decision between building an ad-supported consumer product or a paid-for enterprise product.

Are there no alternatives? After nearly a decade of neglect, Apple recently woke up to developers’ pain and announced sweeping changes to subscription options within the App Store. But is it too little too late for a generation raised to expect software for free?

Perhaps the answer lies in the East, in the examples set by the Asian Unicorns which employ radically different business models to the traditional product playbook of Silicon Valley. Line (which recently floated on the stock exchange), pulled in $1 billion in revenue in 2015 from in-app purchases of its social games, targeted ads via registered accounts and (shock horror) the sale of actual stickers! Yup, that’s physical merchandise, folks. Having accepted that people won’t pay for digital products, they leverage them as channels for selling physical ones instead. This is something ELT publishers have always been wise to, hence everyone backing the blended learning horse rather than the 100% online one – blended learning = book sales + SaaS sales.

The one to watch, though, is WeChat. WeChat is a product phenomenon in many ways, most notably in the way it embodies a new law of digital reality that many are claiming is a perpetual break from the past. The law in question states that products are merely ‘pipelines’ whose time has passed. The future for digital businesses lies in building platforms. Whether this is, in fact, the case is will be the subject of a future post. Watch this space.



Paul JacksonPaul Jackson, long-time friend of ELTjam’s, is a Product Manager who has been building digital products and services for over 15 years. He is currently Director of Product for Newsmart, a digital service that helps professionals master business English with premium news content (which ELTjam also works on). He is also the publisher of Pivot Product Hits, a weekly newsletter on digital Product Management. In this guest post, which first appeared on Paul’s blog and has been adapted for ELTjam, he explores digital business models that might work for ELT.


Image from: https://www.flickr.com/photos/khawkins04/

1 thought on “Why so many digital ELT businesses fail to make money, and what we can do about it”

  1. I appreciated this post, lots more questions than answers …. Agree B2B (or B2S) is the best approach but takes start up capital and time.

    I also think that one answer to the question is simply that the businesses/start ups don’t ask customers to pay right from the start. They don’t show they value their product and start small with customers that will / are paying. That’s where the slippery slope begins. Pre launch, sign up customers to pay at X. A good number of platforms that will allow you to do this and not process payment until launch. Launch and then build off of this paying base.

    I have a fundamental problem with ads in educational products, online or off. Even just visiting blogs full of stupid ads turns me off and I wonder about the generation of teacher leaders we are creating – where sharing their professional knowledge online is seen secondary to and a means for generating paltry income from adwords.


Leave a comment