Audience Notes from Edtech Podcast event on Monday 20 June at NESTA
I was asked to sit on this panel by the delightful Sophie of Edtech Podcast, who briefed us on what we needed to focus on. Being time limited and a tad verbose, I’m always a bit frustrated as I always have far more to say than would be polite to my fellow panelistsl; to compensate I have written a few notes which may be of interest (or not).
My first overarching answers is that edtech isn’t a thing, education is. Edtech by itself is a cul-de-sac, but integrated into the heart of education it becomes part of a complex and vitally important puzzle.
Q1 How should we value edetch?
3 main perspectives:
1 -Educational (pedagogical)
- Measuring edtech in terms of efficacy and pedagogy is almost futile as the arguments and ‘solutions’ are ideological, political or in the case of efficacy, whatever you want them to be.
- There is a significant disconnect between academic research and the realities of education, e.g. it takes too long (up to 7 years), decision-makers and users are bad at sifting the pointless from the worthwhile. Testing is a great example of this disconnect. Teacher unions and “jobsworths” all rail against SATs and similar, then cry out for more money and reform when we get poor results from PISA and other external reviewers.
- Two initiatives I like in this area are research are reserachED and the swift research models developed by Rowland Fryer’s team at Harvard Education Innovation Labs
- Governments are crap at education because they are driven by ideology not pedagogy. Most Secretaries of State of Education either know nothing about education or are in place too briefly. Even the experienced civil servants in the endlessly renamed DOE are often blockers – sometimes for personal political reasons, often as part of institutional group think, or as as they navigate their way around the civil service behemoth.
- 9xSoS since 2001 = average tenure 1.7 years
- Gove longest = 4 years
- Best in my opinion were Blunkett and Gove
- I worked at DoE for two years, which was instructive but no fun for me or the poor people who had to try and work with me.
- Basic financial metrics allow us to more accurately assess failure, success and the gap in-between.
- There’s loads of financial metrics like EBITDA and Discounted Cash Flow. Personally I’m interested in sales, margins, retention, fixed and variable costs and whether the business can be solvent without needing endless fundraising at stupid valuations.
- Sophie’s M&A analyst friend is wrong – his/her comments show a lack of knowledge of the sector domestically and internationally (education is simultaneously a series of global, national, regional and local markets). For example, Pearson is a UK FTSE 100 company with a market capitalisation of around £6bn with turnover of £4.5bn. Yes, edtech is a subset of this but it’s a growing part which is of vital importance to all the big players, who tend to do edtech badly.
- Most big-ticket educational M&A is transnational, e.g. the $190m Series D round just raised by EverFi in the US, of which $120m came from RISE, the impact fund of TPG who own TES Global and whose investors include Sir Richard Branson, Bono and Lauren Powell-Jobs and a few more
- Impact investing. NESTA have two edtech investments, which I like not because they are impactful (by NESTA’s metrics)l, but because they are fundamentally great businesses led by outstanding leaders (3rd Space Learning and Arbour Education).
Best UK edtech exit?
- Education City
June 2010 £70m 14x EBITDA
- UK’s most successful educational entrepreneur?
Andrew Colin – The Study Group and INTO University Partnerships
- Listed Australian company who paid £24m for 40% of a UK edtech company?
3P Learning and Learnosity
- Most active M&A firm in edtech internationally (by value)
- Listed US company who is an investor in Chegg, CLEVER, Cousera, Knewton, Course Hero & General Assembly?
GSV Capital (who co-own the GSV/ASU conference) IPO’d in 2011 at $15, currently trading at $4.50.
Approximate size of the UK and US for-profit edu sectors?
- Education spend is generally 90%+ public and 7-10% private-for-profit: so of the UK £100bn around £7-£10bn is private; in the US the numbers are $1.4tn and $98-140bn.
- M&A is problematic in edtech because it’s a transactional business that’s like a bad marriage. In the beginning there’s lots of dating, foreplay and goodwill, then almost immediately after the relationship is consummated, one party starts planning how to most profitably exit the relationship within 3 to 7 years (and move immediately on to another).
- You can also value edtech and education in terms of brands and exports. Sadly the brand of UK education PLC is far less robust than many imagine. In absolute terms we will probably lose 2nd place in HE to Australia, who have over 800,000 fee-paying foreign students and a government who tries to understand and values educational exports (something Teresa May never did at the Home Office and seems blind to as Prime Minister). I recently heard Lord Biltmora praising the excellence of UK PLC’s education brand and the British Council. He was very sincere but out of touch and seemed to have no idea that in education the British Council are not seen as a trusted neutral partner but as a poacher and gamekeeper which competes with you from behind the skirts of UK diplomatic service.
- Finally, can you make money and do good (‘impact’ and not-for-profit investing).
- Firstly, many not-for-profits don’t have a commercially sustainable business, which is why it’s where most of the $5bn+ p.a. (in the US) from foundations like Gates/Ford/Carnegie/Lumina/Broad Foundations ends up. Yet even with no expectation of a financial return most have found it very hard to find good educational investments.
- Interestingly, a Harvard Business review Study found no link between corporate social performance and corporate performance. My take on impact is that edtech busineses with a solid financial model, who pay their staff fairly, pay tax and give a reasonable return to investors are the true social ventures, not those that go around with a charitable begging bowl.
Q2 – Navigation of the edtech ecosystem
- Navigating edtech is hard for developers, buyers, users and investors.
- We have had loads of initiatives like E-Learning Credits, BECTA, BBC jam, DfE Centre for Procurement Performance and many others which became bloated bureaucracies and/or snouts feeding at the government’s financial trough. Few have had much direct impact with students, teachers or learning as they are mostly non-strategic (political) and often counterproductive (the mirage of ‘joined up policy’). Even where big budget programs are said to have been successful, like London Challenge, I’m still skeptical as in this case I think the impact of tutoring and the strong educational aspirations of immigrant families are as likely to have been causative factors in London Challenge’s success.
- Having procurement frameworks is a great idea if they focus on things like TCO (total cost of ownership), Opportunity Cost, etc, but aren’t a guarantee of ending the stupidity I continue to see at an organisational and individual decision-making level.
- I believe that all publicly-procured edtech should have Open Data APIs (with anonymised data) so that there can be more independent research to see if they are worthwhile (I’d expect 90%+ aren’t). Proof, efficacy or whatever you call it is tricky and subjective but we can do much better at assessing the opportunity cost of edtech generally, and specific products and services individually.
- Personal and professional recommendations are the king and queen of how edtech is found by teachers. At a tech level Google is the gorilla in the room but the way its personalization technology works isn’t great for edtech. Two edtech companies doing interesting work in this area are TeachPitch and Biblio.
- Efficacy is just what you want to measure today; tomorrow it’s something different. My cynicism about this comes from having heard all this stuff several times, from proselytizers who claim to have found education’s holy grail, only for it to disappear and then reappear as something ‘new, radical; and transformative’. As I get older it seems more obvious that real change in education as in most things is incremental and unsexy. Sexy, big excitement stuff is educational Viagra, great while it lasts but actually just masking serious underlying dysfunction. That’s sort of how I feel about NESTA’s RocketFund, that seems to me to be superficially attractive, but is just fiddling at the edges to help the ‘worried well’.
- In theory great edtech should benefit from the network effect of scale with more users equaling a better product as with Turnitin and Teachers Pay Teachers (TPT). However, this is another over-hyped theory as it has largely failed in other examples, most notably TES Resources, whose management recently trumpeted about their ‘billionth download’. Yet TES Resources has tried the TPT model and failed, and most of their free resources are bloody awful and I’d argue one of the biggest wastes of teacher’s most precious resource, their limited time (covered later under Opportunity Cost.
- One of the biggest issues in edtech is the disconnect between educators and students and how edtech is imagined, developed, sold and assessed. I tried with Cambridge University Assessment to shift this in the pilot program (now closed) caleld ed-invent. To me it’s still a big issue.
- I get vexed about bad research, overstated results and bad reporting. A great example is the coverage of an Irish iPad experiment on the BBC.
- Unfortunately, this isn’t a one-off and I was equally underwhelmed by some research into an educational app funded by the Sutton Trust and undertaken by Oxford University. It was excellent research but the overstating or cherry-picking of results to fit a media cycle discredited both the funder and Oxford as the conclusions as presented to the media weren’t strongly supported by the results.
- In theory bad results should be the death knell for any edtech products and while you could argue there should be correlation, from an investors’ perspective I’ve seen very little causation at least in early stage edtech.
Q3 -‘Protecting value of edtech’
- Budget cuts in UK education are going to be a reality where even a Conservative government pigeon-holed as being all about ‘austerity’ still manage to rack up £50bn of debt each year. The interest on UK debt is already more than we spend on education or defence and if Mr Corbyn and his ilk get in (the impossible now seems possible) then UK PLC will, like in the 1970s be bankrupt and have to go begging to the IMF to bail us out. If you think this isn’t possible it has happened before and the price the IMF demanded was a 20% budget cut in one year, something that makes Tory ‘austerity’ seem like a picnic. So, my view is that budget cuts are inevitable, but they should really be focused at the biggest areas of cost, that is salaries and pensions. The scandal of the unfunded £176bn Teachers Pension Scheme for England and Wales is where I’d start. If Ontario Teachers’ Pension Plan can, with 360,000 members, grow to £90+bn under management and give a 10.5% annual return for over 25 years then it’s possible. As controversial as it seems I’d outsource our local scheme to them, but I’m sure the opportunity cost would be too high!
- The benefit of any cuts to edtech is that it should sharpen the focus of those who will have less and hopefully help kill off some of the cargo cult-type ideas such as iPad’s automatically equate to better educational outcomes. I still hear this sort of crap far too often, and mostly the tiny minority of frontline teachers who are tech enthusiasts (my ‘usual suspects’).
- The best way to protect data is to make it open, anonymised (one example is Assembly from ARK) and owned by schools and even better individual users. It’s not easy and the raft of state-based initiatives in the US is dizzying. Local edtech companies will have to comply with EU’s 2018 “General Data Protection” legislation but for many, the US and other markets are far more important when it comes to data. Any investor who thinks the siloed data in an edtech product is their gold dust is deluded. Data should be open and if it doesn’t show the product or service actually helps teachers teach or students learn, then you‘ve simply backed the wrong horse and no amount of PR should be able to hide this fact.
- There is NO SUCH THING AS FREE and educators, students, parents and politicians need to wake up. If it’s “free”, then you are the product or it’s being funded by philanthropy or sponsorship. The latter may seem as bad, but it depends on the sponsor. In 1997 I was a director of a company that created DaisyMath. We gave it away for free because I got Kraft to sponsor it and they didn’t want us to give them any data except to show it was being used. This isn’t to say they didn’t evaluate the deal, they did so ruthlessly but against a series of brand metrics that were all predicated on did it help ‘make a difference to the lives of children and families’ (their customers). They also looked at whether it had any impact on how their employees and distributors (shops and supermarkets) felt about Kraft. Measurement of this was outsourced to experienced research and media agencies and if it didn’t hit their targets they’d have dropped it immediately (they tracked this stuff 24×7). Fortunately, it was, at that time, by Kraft’s own assessment, the best sponsorship they’d been involved with and lasted for 4 years.
- Too many edtech products have a free version. If you want to know why this is a big mistake go to Edsurge and read the story about this very issue by the founder of edublogs.
- It’s hard to create or protect value in edtech when too many schools and students have poor bandwidth. A US survey a few years ago showed some schools paying 14 times more for the same bandwidth and almost 75% had less than 100mb per 1000 students, the key benchmark of the US State Education Technology Directors Association.
- Protecting edtech also strays into the arguments about teacher replacement. My view is that in places where kids have no school or terrible teachers, then tech is a huge benefit. In the UK the biggest threat is the failure to deliver after three decades and £20bn, yet we still have a minority of teachers using it well. Controversially I think there is a need to stop giving educators lots of carrots to use edtech and to start using a performance management stick. However, this is a two-edged sword and teachers reciprocally would/should demand they are only give access to great edtech.
- A race to the bottom financially exists both in startups, established players and amongst the behemoths like Google, Apple, Amazon, Facebook and Microsoft. The ‘Frightful Five’ as the New York Times labelled them, can afford to give away their products and services for free or heavily discounted in education as they see it as simply a cost of doing business with government. It’s also cheap reputation-polishing to deflect regulation and punishment for their more egregious activities like avoiding tax. Personally I think the most substantive thing the Frightful Five could do for education is pay a fair share of tax in all the countries where they operate. A dream perhaps, but if they continue to rewrite or avoid the basic contract between enterprise and the state (pay a fair share of tax as the license fee to operate) then we will all suffer in the long run.
Q4 – Opportunity Cost
- Rarely discussed in edtech by companies or users, it’s hugely important for everyone involved in edtech and education.
- “Opportunity cost” comes from classical economics and is tied up with concepts like relative scarcity. It boils down to what is the thing you give up to do x? That could be wasting time searching for or creating free resources, a hobby John Hattie correctly identifies as being pervasive in education. At a practical level it’s why educator and behaviour expert Tom Bennett is critical of the uncritical embrace of edtech and particularly Minecraft. When he said it may not be great for schools and students he was howled down online by Minecraft enthusiasts, teachers, students and almost everyone. But what he said was exactly right, Minecraft may have great educational benefits, but in choosing to use it we have to make hard-headed decisions about what we drop give that budgets are fixed or declining and that the school day and curriculum is already crowded. Is what goes; art, a language, a teaching assistant…?,
- The most important issue relating to opportunity cost and edtech relates to the most valuable resource in education, a teacher’s time. The difference between the use of TES Resources and TPT is an example. A teacher’s time is worth at least £40 pa and so 2 hours trying to find free resources comes with an £80 opportunity cost. The advantage of TPT is that there’s simply far better content (primarily for US teachers) at TPT because:
a) it’s a meritocracy
b) the network effect means more buyers attracts more good content from educators
- Ironically, while I’m not a fan of TES Resources there is a local example that shows schools can do this as well as anyone. That example was Thomas Telford Online, an edtech business created by a school that went on to earn over £12m profit, all of which was reinvested in the local education community.
- I spoke to Tom Bennett about the issue of giving teachers time management training and his view was that more needed is:
– better leadership training to not over-task staff
– better training for teachers to help them differentiate the meaningful from meaningless in their class and schools
Q5 (my questions at last)
- Navitas’s report is irrelevant in K-12 as they are primarily FE and HE. This year they launched edugrowth, an edtech incubator, where the first cohort was in my judgement pretty lamentable.
- If you want to look at the reality now and future trends in edtech, one of the best sources is the annual NMC Horizon Reports.
- Edtech in particular and education in general also needs to get a hell of a lot better at business intelligence (historical, vertical and horizontal and current). Having founded and still part-owner of The Assignment Report, I think it’s the best starting point in the UK.
- Edtech has asymmetric warfare where bigger companies, experienced investors, funds and private offices all have better lawyers than most edtech companies. If you are a great edtech or tech company in general the best defence is to pay a seriously experienced (expensive) tech lawyer to draw up a term sheet that says what you want and to say to investors, ‘here it is and if you don’t like it bugger off and stop wasting my time’.
- Edtech startups also need to get better at strategic planning for fundraising and avoiding the common ‘oh shit we’re running out of cash’-crisis approach.
- I think roll-ups are a big opportunity in early stage edtech. Locally we have seen this done by Big Clever Learning, LTG and Sandbox & Co, but for behemoths the action is in divestment, e.g. Pearson’s sale of the Family Education Network to Sandbox & Co and Scholastic’s sale of its entire edtech portfolio to Houghton Mifflin Harcourt for $575m in 2015.
- Getting a return in edtech is about surviving the race. It takes 7 years with most horses falling and having to be put down before the year 3 fence.
- Most edtech startups have stupid valuations generally of a minimum of £1m when for the same amount I could buy 100 new Kumon franchises or a small McDonalds.
- The dirty secret of edtech is the difficulty is exiting. Private companies are a generally illiquid market, where in many cases you are a price taker and a price maker only in the very rare event of a huge success. MoshiMonsters is my £180m painful example. Paul Birch and I sold to MindCandy (the company behind Moshi) our edtech games startup Tutpup for roughly1% of their equity. It was early days and as the brand grew to 90m users so the valuation hit about £180m. Valuations of this type by analysts are complete bullshit as they are primarily built around a financial model known as Discounted Cash, something accurately described as ‘a rough guess multiplied by a wild estimate’. At this time I was offered a princely £60k for my 0.5% that were supposedly worth £900k. This was pretty galling although in retrospect I should have taken it as now the value I place on those same shares is £0; I do however have the share certificate framed and ready to hang in my toilet!
Best and worst deals
PowerSchool founded 1997
- Sold to Apple for $62m (all stock) in March 2001
- Sold to Pearson $40m May 2006
- Sold to Vista Equity Partners (along with other LMS systems) for $350m June 2015 (+6 rollups)
- If Greg Porter and John Stone had kept their Apple shares then they would have earned $231m in and the shares would be worth about $6bn giving a total of $6.2bn (the same as Pearson’s mkt cap)
- 4 founders
- No investment
- Kraft sponsorship $300k cash +$1m co-op marketing
- 4 yrs $1m ROI to founders
- We all withdrew and left the founder who never made another cent from it
Night Zookeeper (personal)
- Picked as winner of Learning Without Frontiers startup competition. Of top 3 the only one still in business
- 1st investor
- New deal with major international publisher
- Australian ISP that was sold into the Australian education market by Reed Elsevier in 1994. Ozemail’s education business was a key platform for its subsequent growth that made Malcolm Turnbull, now Australia’s Prime Minister, a multi millionaire!
Aren’t the ones I lost money onm, they are the great edtech investments I passed on or missed, like:
- Amplify – News Corp dropped $1bn. Remnants acquierd by Lauren Powell Jobs via the Ememrson Collective, where they rolled up Amplify Games with Story Toys (Ireland) and TouchPress (London) under the TouchPress brand
- Riverdeep takeover of Houghton Mifflin now Houghton Mifflin Harcourt
NYSE Houghton sold to Vivendi $2.2bn
On sold to Bain Capital & Thomas H Lee for $1.7bn expand with a lot of debt and struggled to find an exit so appointed Goldman to do an IPO to be outbid by Barry O’Callaghan’s $3.3bn Riverdeep (tiny/failing Irish elearning company) who then acquired Harcourt for $3.7bn (Houghton Mifflin Harcourt) to bankruptcy in 2012 Nov 2013 share IPO valued c/o at $1.7bn same as 10 years before