On my first day at ELTjam, I was pitched at. Tim, Laurie and Nick (the ELTjam founders) sat me down in the meeting room at the Hackney Wick co-working space where myself, Nick and Hristina are based, and they tried to sell me their product ideas. This week, we take the first step towards testing whether one of these ideas can become a viable ELT product. Today sees the launch of the flovo.co landing page, our own MVP, a little litmus test.
But how did we get to this point, what have we, specifically I, been up to since the 12th May and why have we made a webpage before the app is ready? This post explains the process that we’ve been through and some of the decisions we’ve made. We wanted to share some of the things that we are learning in the hope that it can help others involved in both ELT and in EdTech, and we also wanted to get thoughts and ideas on how we can improve the process we’re going through as we create digital ELT products here at ELTjam.
A business model canvas encourages business owners to properly think through the various aspects of a product or idea before investing more time and money in making it a reality. In 2009, Ash Maurya created an updated version, the Lean Canvas, reflecting the move towards leaner working processes. The principles are the same; put simply, you think about who might use or buy your product, what problems they have and how you could solve them. From this you write a value proposition (what benefit you provide, to who and why those people would choose your product over any others available). You then look into how those people would be able to find out about your product or service and how you’d make enough money to make a viable and sustainable business out of it. It can be a confusing and time consuming process and very often leads to a realisation that the great idea you had has very little chance of working or significant obstacles in its way.
In my first week I worked on, or looked at, initial lean canvases for around seven different products and services. The idea is that the ones that seem viable get added to and built up over time, the ones that are problematic need a change of course, a pivot, in order to get them back on track, or they can be scrapped. By the end of the first week there were four product ideas that ELTjam decided to pursue (three of which are still being actively worked on), this blogpost tells the story of just one, now known as flovo.co.
The product idea
The product is based on three main assumptions: The first was that learners need vocab, and lots of it. If learners can get a foundation of the top 3000 words under their belt, the rest of the language learning process will be smoother and more productive. The second was that learners needed to know more than just meaning, they needed collocation, pronunciation, understanding of word families, synonyms, range etc. too. The third was that if we harness the power of ‘flow’ (the feeling you get when you’re totally focused on what you’re doing, you’re ‘in the zone’), more learners would be able to get through that volume of content and more quickly.
Value proposition: flovo.co is a series of games that help users learn the words that matter (high frequency), learn the things about these words which matter (appropriate depth), and learn in a way that maximises chances of reaching a state of flow (highly motivating).
We figured that enough learners knew that there was a need for vocab, and enough teachers too, and that the market was therefore there. In terms of business model, we felt that there were two possible routes, either direct to consumer (B2C) or a business to business model (B2B). These two routes vary greatly in terms of channels through which people hear about the product and the ways that they generate revenue. We looked at both of these options and decided that there was enough potential to take the next step with the product.
The next step was to work out what assumptions we were making with this product idea, and which ones were the most risky; it’s best to start by testing them first! We therefore had to research in more detail the different things that we looked at in the canvas and think about what our most risky assumptions were and go about testing them. The book Lean Analytics by Croll and Yoskovitz says that it’s useful to think about knowledge, or information as falling into one of four categories:
- Things you know you know
- Things you know you don’t know
- Things you don’t know you know
- Things you don’t know that you don’t know
Clearly the biggest risks, but also the biggest opportunities, are in the unknown unknowns. Data can be gathered about the known unknowns, but early stages of product development should be about finding out as much as possible and in doing so you may stumble across something, an opportunity or an obstacle that that you hadn’t even been looking for in the first place. This will generally result in a pivot to a new product direction, revenue stream, channel or proposition. The problem with coming up with an idea and then just building it is that there may well be, in fact almost definitely are, things you don’t know and have no idea about and they could and should have a big impact on how the product progresses.
So we had to decide what to test and how to go about it. We figured that our first main assumption is that people would be interested in a product as described above. Do learners care about vocab or is grammar more important, do teachers and parents want learners to sit there and learn a few thousand words in isolation? Would people be willing to pay for such a product even if it existed?
The way we decided to test these assumptions is with a technique, now not at all uncommon, whereby a web page is built that highlights the value proposition and asks people to give some information to register their interest. By building just a single page site we’re hoping to be able to test some of the assumptions that we’re making, measure people’s responses and gauge interest. The things that we learn from this stage will help us to further define how the product should look, which markets it should be aimed at and the types of people that are interested.
Building the page
It was tricky, at this early stage to decide how much time and money we should put into different aspects of the page. Do we pay a web developer or just build it ourselves in WordPress? Do we pay a designer or mock something up as best we can? Is it worth getting someone to do copy, or is that a place we can save money? The problem is that there is a certain standard of website that consumers have come to expect, and no matter how good the value proposition, if the design is poor, the site clunky or the copy unappealing then people will be less likely to engage. But at the same time we had to keep costs as low as possible so it was a tough decision to make. I spoke to designers and developers from different countries and in the end decided to work with Mark Bain for this stage of the product. He had experience in working with other ELT products, could do both the development and the design, we liked his work, and he quoted a price we could afford to pay!
Then started a process of back and forth between Mark and myself as we iterated on design, logo, layout, copy, functionality etc. We wanted some images of app screens to give the impression of what the app could do, so we spent some time discussing functionality and how best to show the features of the product, despite not being 100% sure of what they would be! We started with overall themes and impressions that we wanted to give and then designed screens and wrote copy to match the ideas. It was also really important that the website was responsive so that it would look good on a mobile, a tablet and various sizes of computer screen, and in order to store the email addresses and create the signup forms we integrated our mailchimp account. But most importantly, we also had to decide on a name. The working title was Flowjam3000, which still has a special place in our hearts, but once a member of the ELTjam pointed out that it sounds like a different product altogether, we realised it needed to change! We created a spreadsheet, and all wrote down a load of terrible and a few good names, and after a lot of domain name searching we settled on the achingly cool flovo.co!
SEO / SEM
As a general rule, in order to get people to your site you can optimise the page to rank higher for certain searched terms (SEO) and you can pay to have your page appear when people search a certain term (SEM). Basically, it’s a market that Google have got all sewn up! You can advertise in other channels such as facebook and twitter too, but at some point you’ll have to put some cash google’s way!
In order to optimise a page, you need to include high quality and relevant content for the people searching the phrases you are interested in. This is therefore much more of a long game, and one that doesn’t really fit with a single page website. You need a blog, info pages and various parts of the site to populate with relevant content. For this reason, we felt there was little we could do in terms of optimisation with the single page we had, and we didn’t have enough time and money to get quality content before we’d validated whether the idea had potential.
So we turned our attention to SEM. How many people would view our site for £10? And how many for £20? It all depended on which words we bid on. So do we bid for big terms like ‘learn English’ or go niche with phrases like ‘English vocabulary app’? What would we learn from paying this cash? How would it influence our decisions? Do we do it in English or Spanish? So many questions, and the answers to each just threw up another few things that we didn’t have enough information or knowledge about. So we decided on a slightly alternative approach; first we’d release the page in English and get feedback from our peers and a few focus groups of learners. From what we learn we can decide whether the idea has enough potential to translate everything to another language and then spend some money on advertising in that language and see if there is genuine consumer interest. Hence the first version of the page, the one you can see at flovo.co is all English and not graded for language learners.
Building the app
Whilst the page was being built, we started looking into other areas that were risky for us. Firstly was the question of whether or not we could afford to build the app, even if people did want it. We had contact with a company called Zzish, who provide back end modules for developers of educational apps. Their idea is to reduce development costs by making available the modules that are common to a large percentage of educational apps and tools. So instead of every company spending money on a gamification module, a learning analytics module, adaptive algorithms etc. Zzish can provide you access to these modules through their API. After approaching Zzish it seemed that there was a big overlap between the functionality they offered and what we were aiming for. They also seemed really interested in the project and so quoted to build the whole app as well as integrate their own modules through their APIs. Zzish have a great team in place and do some really exciting work, plus the price was affordable. So assuming that we validate the idea, things are in place to build the app, hopefully within a month from getting the green light.
In order to develop the games and exercises in the app, we need a database of words to build upon. We need definitions, example sentences, derived forms of headwords, links between words of the same word family, words that share a sense, words that have opposite meanings etc. Doing this ourselves through a process of natural language processing would be theoretically possible, but very time consuming. We expect that data that the app would generate would hopefully be of some value in terms of finding out more about how people learn different aspects of words and any regional and age variations etc. So we approached a publishing company and we’re in talks about how we might be able to use the dictionary data that they have available and what implications there may be for the data that we collect. Things are still in early stages but the hope is that this dictionary data will form the main content layer for the app.
Games, progression and scoring
We’ve also been putting effort into what the games will look like, how the adaptive algorithms will work and exactly how we’ll manage to achieve a mental state of flow for our users. Again, it’s not something we want to spend a huge amount of time working on, until we know that there is interest on the product. But as it’s the exciting bit, we haven’t been able to help ourselves. We’ve got some exciting ideas, watch this space!
A lot has happened since that first pitch day here in Hackney Wick, but this really is only step 1. Over the next few weeks we’ll be refining the copy and the offering before translating the site into another language, probably Spanish. We’ll then make decisions about how much to spend on advertising and which search terms would be most useful for us. Hopefully, by early August we’ll know if there’s enough interest to get the app built and then we’ll start testing with a small number of users.
So if you, or some learners you know, are interested in being part of the initial group of people testing the app, sign up though the flovo.co page we’ll keep you updated on the progress.