Thursday, December 27, 2012

Why Blended Charter Schools are "Whales" in the ed institutional context...

My past few posts, I've painted a pretty grim picture of what it takes to sell to schools and districts.  I want to take a moment just to answer the question "But hey, Rocketship doesn't act that way, you work with a ton of companies, you pay for products, etc."  It's actually taken us a surprisingly long time to figure out just how different we are.

The fundamental reason we are different is that we are completely comfortable with the idea that we will use the best method for every type of learning for every student.  What this nets out to for edtech is that if there is a program on which students learn certain concepts faster than with a tutor or from a teacher, we are overjoyed for them to learn that online.  The efficiency of online learning (i.e. its cheaper per minute of learning time) allows us to redirect resources to pay teachers more, support them better, and put even more effort into the code we write to understand our students better.  Because we are incredibly incented to allow students to learn as efficiently and effectively as possible, every time we can use edtech, the cost of that program is insignificant on a learning per minute basis compared to our other options.

So while traditional schools often having no budget line for edtech and having to steal money from their curricula line item, any serious blended school has a budget line for edtech.  Ours is $100/student.  Before your mouth starts watering too much, that $100 gets cut up among dozens of programs for assessment, instruction, scheduling, etc.  So it's not a goldmine, but the point is that we have a line item.  That means when you come to us with something good and we test it and it works, we want to buy it just as much as you want to sell it.

Compare this behavior with traditional schools.  Traditional schools are not agnostic about learning methods.  In general, they believe that a teacher with their class of students is the one and only best learning approach.  If they happen to have computers in that classroom, the number of kids learning on those computers has nothing to do with the number of students that teacher can impact.  So economically at least, they are disincented from purchasing technology.  The school has already sunk the cost on its teacher salaries, and providing better online learning might be nice for the teachers, but since the perspective is that learning comes only by teachers teaching, the technology is just an afterthought.

So that's basically how blended schools are different.  For many years, we implored vendors to build things for us and we convinced quite a few.  What we've realized is that we aren't a very big market yet.  Even within the charter sector, blended schools have only started to take off, there are probably fewer than 100 truly blended charters at this point.  And on the district side, I have not seen a district make labor trade-offs yet, meaning that technology is not strategic, just a support for the teacher doing everything themselves.  So a market of 100 schools with maybe 50,000 students is not nearly big enough to build a company around.  How quickly does this change?  My guess is that within the charter sector, this begins to change extremely rapidly (for our world), but nowhere near fast enough to worry about for the next decade.  I would not be surprised to see 1000 blended schools in a decade, representing 500,000 students.  If you had a great product and could charge $5/student, that would be $2.5M in revenue.  Not so good.  Outside of the charter market, is there a chance that things flip over quickly to blended and we see edtech become strategic?  I'm usually pretty optimistic that change can happen, but unfortunately this goes right at one of the sacred cows in education - that only a teacher can provide the learning a child needs.  There are many forces between unions and districts and school boards that will be very reluctant to rethink this belief. Even when Rocketship and other blended schools are getting 2X the learning for kids and 2X the salaries and job satisfaction for teachers, I think those institutional changes at traditional schools in district will be painful and slow.

Higher ed is a really good analogy here.  The new wave of online higher ed companies like Udacity are going directly to students (for free!).  While there may be some universities that get it and will be able to address the needs of students as well with a blended model (EdX from MIT and Harvard for example), the vast majority will not make the transition and there will be a new way that students learn after high school.  This undoubtedly will make access to quality courses much higher and create a lot more equity than anything we could do physically (absolutely everyone is going to have a smartphone before the traditional institutions get on board).

So as a K12 edtech entrepreneur, I think that if you make the student your customer, you are in good shape.  Schools like Rocketship will be great ways to validate and test products and we are happy to do it, but ultimately, you are going to make your money by doing a great job with the students themselves and we are just a channel to the student.

Monday, December 17, 2012

My Trip to India Part 4 - Affordable Schools

It was very interesting to be in a country where over 50% of students attend non-government schools.  Let me say that in my opinion this proliferation of school choice is probably the second best reason to have hope for India, right behind the fact that every Indian parent is willing to do whatever it takes for their children to get a great education (so they can care for them in their old age).  Now let's talk about what it may take for the affordable schools (generally with tuitions of $4-$20/month) to become a major driver for improvement in the system.  For a long time, charter school advocates have focused on getting large penetration rates.  We have found in places like New Orleans (80%+) and Washington D.C. (50%) that systems can start to make profound improvements in student achievement with high penetration rates.  It's important to unpack the conditions that must be necessary for this to happen though:

1 - There must be a strong accountability system which can differentiate and compare the quality of all schools in the system.  This is true in the United States, but not true in India.  Before NCLB, we had many charter schools started for non-academic means.  Like India, they were generally regulated on inputs/process, but not on outcomes.  What we found is that they didn't do any better than most schools on average, because many of them were not started to produce superior academic results.  I would guess the exact same thing is happening in India.  Because the schools of choice are not accountable for student results, and because they are run as small businesses, they are going to maximize profits at the expense of student results.  So once India has an accountability system, will really be day one of creating a great school system.  Just as in the U.S., they will probably find that 10% of schools are doing incredible work and need to be scaled up, some other set are doing no harm, and a bunch need to be closed.

2 - It must be politically and legally possible to close bad operators.  We see this difference between Washington D.C. and New Orleans.  D.C. for a lot of reasons has not had the will to close down bad operators, many of whom are incredibly well connected politically.  New Orleans on the other hand has had a much clearer policy on academic results schools need to show.  They won an i3 grant a couple of years ago to begin shutting down the bottom 10% of schools every year and re-opening them with high performing operators.  Rocketship's charters were approved in New Orleans under this program.  So in India, the question is how well the market will work on its own and how much the government will have to get involved.  I would imagine that just as in the U.S., shutting down government schools will be very difficult.  I wonder though if the affordable schools will actually face market closure by parents who don't like their results.  We have found in the U.S. that parents choose schools for a lot of reasons - proximity, safety, special programs, etc.  My suspicion is that outcomes will be a lot higher on that list for Indian parents, but we will see.

3 - Talent.  The challenge with building a great school system always comes down to how many talented educators you have and how well you can leverage them.  The U.S. charter system took decades to work until there became enough high-quality operators that could scale up.  Once a few of those exist in a city, then a city can decide how many great schools it wants.  We are probably still a decade away from having a large enough supply of quality charter operators who can scale at the rates necessary.  (Politically, you have to make a big dent in the city's problem within 5 years in order to create the political will to finish over a decade or two)  India probably does not have anywhere near the level of talented multi-school operators it needs, though this will not become clear until the accountability system is in place and we can see who gets the results.  If the U.S. is informative, until we had NCLB, we had almost no "No Excuses" schools in the country.  So I would say again that once the accountability law has been in place for a year or two, then the country will be able to start scaling up what works.

4 - Political Will.  This one is really interesting in India because it's currently in a very different position than the U.S.  Here, over 80% of kids go to government schools.  There, it's closer to 40%.  So the real question in India is whether reformers can keep the space deregulated long enough to get items 1-3 going.  If they do, they will have an amazing advantage over the U.S. and will likely be able to leapfrog our results, because they won't be fighting to fix a broken system, just focused on creating a new one.  Given the urgency that Indian families feel towards education, my sense is that there will be far more popular desire for choice to continue to proliferate, but parents must be organized in order for this voice to be heard.  I saw no indication that anyone was organizing parents in India and hope that they can get their soon.  A Stand for Children India would be a phenomenal thing to have standing right next to Teach for India!

In the last part of this series, I'll talk a little bit about technology in schools and homes in India.

Saturday, December 15, 2012

My Trip to India Part 3 - Accountability

For a country as intentional as India, it was pretty incredible to see the lack of outcome data.  Students take a test once in 10th grade, and that's pretty much it.  It caused me to reflect on the value of No Child Left Behind.  There aren't too many things I agreed on with George W. Bush, but this was just an amazing step forward in accountability in the United States.  I remember soon after that law got going, I presented my first charter application for Rocketship.  Part of that process was to justify the need for a new school.  So I took the school district results and disaggregated them into the results for upper-income students and low-income students, for white students and students of color.  What I showed was a 50 percentage point gap in the number of students who were at grade level between upper-income and lower-income students.  As I was presenting this to the school board, the superintendent's face got redder and redder, and he was clearly incensed with me.  I could not figure out what was going on.  I asked one of the board members afterwards and her response was telling: "John, we have never seen those numbers before.  Our average result across all students was fine, so we thought our district was doing well."  This was a school board member!  Without outcome data, it's simply impossible for us humans to figure out the size of a problem, we aren't equipped to do that by observation.

So here is India with 400 million kids to educate and this incredibly rich system of government and private schools doing the work.  Within hours, you are talking to people about why a given school approach is good or bad, all completely without any data to prove effectiveness.  In the vacuum that this lack of accountability creates, the government of course regulates the only way it can, with inputs and process.  So India is going to have to move very quickly to get outcome measures in place.  My guess is that they can start with a couple of grade levels with beginning and end of year tests so that they can see gains, probably 4th and 7th grade or something like that.  They also have the complexity that we do, which is that private schools could easily decide to opt out of the accountability system so that their results were not comparable to the government schools.  For the upper-income private schools, I would expect them to do exactly that.  But for most students,  I think India has a great advantage over the United States.  Because their affordable private schools are small businesses, they are incredibly competitive in recruiting students.  Given the urgency of Indian parents, it is likely that outcome data would become one of the chief criteria they use for choosing schools.  So affordable schools would be highly incented to take the test and then recruit with their results.

This aspect of India's system actually made me appreciate the U.S. approach, by focusing on outcomes (although of course people always question their validity), we have moved to a much more rational discussion than only a decade or two ago.  Let's hope that India makes this jump soon!

Thursday, December 13, 2012

My Trip to India Part 2 - Scale

As I mentioned in my first post, it doesn't take long to realize that India has a scale problem that dwarfs the United States in its quest to have a world-class education system.  Let's do some simple math to start.  There are about 400 million kids in India.  There are currently about 10 million teachers.  If you talk to major urban superintendents in the United States, they will tell you that about 10% of their teachers make a year or more of progress with students each year.  I know that is a low bar, kind of like "do no harm" more than success, but if you could create a school system with this 10% of teachers, you should be able to deliver a very high number of students to college readiness by the time they leave the 12th grade.  In the U.S., this means that we probably have about 400,000 of our 4 million teachers who are able to get the results we need.  We have about 60 million students, so the ratio of students to proficient teachers is 150:1.  In India by contrast, if we assume (probably generously) that 10% of their teachers are proficient, then they should have about 1 million teachers for their 400 million students, so about 400:1.

In the past, we've always looked at a one teacher per classroom model.  In that frame, these ratios don't make any sense.  You are left with the idea that our only path to success is to increase the number of proficient teachers by 10x so that we can get the ratio into the range of a class, say 30:1 or 35:1.  The challenge with this thinking is that massively increasing the number of proficient teachers is incredibly difficult.  There are two fundamental problems.  First, teaching is a low-pay and low-respect profession.  Believe me, I don't like that at all, Rocketship is changing that within our own system, but it is unfortunately true in the U.S. and almost every country in the world right now.  That means that your incoming pool of teachers tends to be filtered to people for whom the salary and respect are a good fit with their skills.  As we have seen, most teachers come from the bottom quartile of academic achievement on admissions tests and college.  Teach for America is the outlier here.  Through an amazing brand and recruiting, it has successfully reached the top quartile of American students to become teachers.  This is just a difficult thing to scale up to 4 million teachers.

The second challenge with creating more proficient teachers, is that improving teaching skill is very difficult.  We have been successful at Rocketship in increasing teacher proficiency by .1 years of average student gain each year they are with us, but at 10 times the effort and expense that government schools spend on it.  We have dedicated academic deans focused on improving teacher performance, our teacher pay aligns with student performance, teachers only come to Rocketship if they want to improve their teaching, we set 3 improvement goals between teacher and dean every 8 weeks, we videotape teachers instructing and compare their work to a master teacher doing the same thing, we use real time coaching with earbuds and live suggestions from our deans.    So the idea that government school systems are going to align incentives and put this amount of work into helping teachers improve is not realistic.  By the way, all of this work is funded by the fact that Rocketship runs blended schools which require fewer teachers than traditional schools.  We reinvest the dollars we save to do this level of talent development, pay teachers more, spend 3 years developing our school leaders, etc.  So without a major change in the way government schools are organized, I'm not sure they could afford to do this level of talent development even if all of the other incentives were aligned.

Given all of this, I think it is safer to assume that we have a fixed pool of proficient teaching talent both in the U.S. and India.  This makes those ratios at the beginning a lot more relevant.  If you have a fixed teacher pool and a given student population, you have to figure out how to leverage that talent over the larger number of students.  At Rocketship, we are moving to a model where our lead teachers are in charge of an entire grade level of 100-120 students.  Of course, they are not the only one in the room with the students, they have assistant teachers, coaches, technology, etc.  In fact, it's much easier to think of the model in terms of the way doctors work in a hospital.  Almost all of the time you are receiving care, nurses, technicians or others are working with you.  The doctor will come in for 2-3 minutes to poke you and ask you questions, then tell the team what tests to do.  After the tests are done, they come back for 2-3 minutes to read the results and make a decision.  So in the several hours you are there, you get about 5 minutes of the doctor's time and at least in my experience, most of the time you get the right outcome.  That is the way I think our schools will work within the next couple of decades. A very highly paid teacher's time will be leveraged in the same way that a doctor's time is now.

For India, that means that instead of that 150:1 ratio we need in the U.S., their proficient teachers need the support structure to work with 400 students.  I am sure they will find many organizational schemes to make that happen, but I think a few things have to be true to get this kind of leverage.  First, the teacher needs to be very well paid, because you really can't ever afford for them to leave their job, they are the linchpin to the entire system.  At Rocketship, we currently pay our top performing teachers about $90,000 per year (about 1.5 times surrounding districts) and will move that to $120,000 per year over the next few years (about double surrounding districts).  Talent is simply too valuable and too scarce in leveraged models to allow churn.  Second, teachers will not be valued for the raw amount of work they can do, but for their decision making abilities.  Figuring out that a student needs to work on a given skill (like a doctor figuring out that you need a certain medicine) will be highly valued.  I am very hopeful that top teachers will continue to show the kind of emotional intelligence that we see today, although I have to say that I think they show a lot more than the average doctor, since doctors bedside manner is less valuable in a high stakes situation than whether they make you well.  Third, there will clearly be other people working in schools with kids than just teachers.  I think that's one place we will see much more selection for emotional intelligence, caring, and a desire to help.  Much like nurses in hospitals, there is nothing like having someone on your side when you are going through a difficult situation.  I think the same will be true at schools.  The good news is that in these highly leveraged models, these folks should be paid about the same as teachers do today.  Finally, a lot of a student's time is going to have to be online, learning skills without the need for any adult intervention.  This is a lot like when you go home and take the medicine the doctor has prescribed.  There is really no need for someone to call you every five minutes to see how you are doing.  If it's not working, you will let them know!  I will be fascinated to see if India can reframe the challenge from the traditional "we just need more great teachers" to one of "how do we leverage our fixed pool of great teachers to serve more children?"

In my next piece, I'll talk more about what India's accountability issues mean for their development, and a perspective that gives on the U.S. system.

Monday, December 10, 2012

My Trip to India - Part 1 - Overall Impressions

I spent last week in Delhi and Mumbai with the generous help of Ashish Dhawan and his staff at the Central Square Foundation, a foundation totally focused on helping India's education system become world-class.  I am going to write about the experience in a few parts over the next week or two, but my general theme would be "What can India teach us about educational systems and what perspective can it provide on what we are doing in the U.S.?"

The first thing you are struck by India is just the sheer number of people.  This is a country with 1.2 billion people, over half living outside the cities, and about half of the population in the cities living at a very low standard of living ($100/month or less).  There are 400 million kids under 18, of which currently only 7% go to college.  So the need for scalable solutions in education is pretty obvious right from the start.

The second thing that is obvious when you talk to people is that India is moving from a developing nation to a developed nation in an incredibly short period of time (a decade or two).  This is very apparent when you are in Delhi (the capital) and you hear about the incredible amount of legislation being enacted to try to bring a largely unregulated society towards equity and the rule of law.

The third thing that doesn't take long to realize is that Indians feel far differently about education than Americans.  Traditionally, Indian children take care of their parents when they grow older.  That means that the fate of the parent and the fate of the child are intertwined for their entire lifetimes.  Sensibly, parents who want to be well cared for know that if their kids get a great education, they are more likely to be able to provide for them in their old age.  So Indians are willing to spend significantly more on their children's education than Americans.  This is not just the top 10% of Indians sending their kids to elite private schools.  The folks I met in this class acted almost exactly like Americans, who also send their kids to elite private schools.  What was much more striking was how the lower 90% of Indians deal with education.

India has government schools just like America.  Like America, there are good ones and bad ones, but the overall average is not great.  They have significant teacher absences, motivation problems, etc.  No big surprise for those of us who have spent decades in government school systems.  What is surprising is the reaction of these highly motivated lower and middle income Indian families.  In walking the streets and talking to people, I found that almost everyone wants to send their children to private schools.  Middle class in India is not like middle class in the United States.  You often find people working as domestic workers, making $100-$300 per month, but they would be considered lower middle class. (average per capita GDP is $1500/yr or $125/mo) These families are willing to send their children to what is called "affordable private schools."  By affordable, they mean schools costing on average from $4-$20 per month.  What is even more striking is that people that don't even make the lower middle-class cut (I talked to several families making in the range of $40/month) are willing to spend the $4 per month for each of their children to send them to private school.  The net of this is that in the major urban areas, a majority (60% was the stat I heard) of the kids go to affordable private schools.  Having a population that has this kind of willingness to sacrifice for their kids to get a great educations is fundamentally different than our culture in the U.S. and bodes very well for a country that literally has to pull itself up by its boot-straps.

The next thing that struck me is that India lacks any accountability system for its schools and teachers.  Kids take one high-stakes test in 10th grade for admissions into what we would call junior college in 11-12th grades.  Other than that, no one has any idea who is doing good work and who is not, except by observation.  As you might imagine, that causes all kinds of dysfunction, because schools which try to innovate are criticized for their lack of conformity, without any way to prove that their results are better.  The affordable schools in particular suffer in this world of unaccountability, because of course they don't look like schools should.  They have awful facilities, poorly trained teachers and enormous class sizes.  So without any accountability system, it's a reasonable reaction to say they should all be shut down!

Finally, just as in the United States, lawmakers really like to make laws!  In India, a new law called the Right to Education act makes the first pass at trying to provide equal opportunity for all Indians in school.  Unfortunately, as we saw in the United States, governments love to regulate inputs and process, while outputs are scary because of the accountability aspects.  So with all good intentions, the government has started to mandate things like what kind of school facility a student should have, the credentials of teachers, class size, etc. with no mention of student learning.  We have seen this play before in the United States.  These kind of input/process regulations show no correlation with student success, but are very effective at driving up costs or driving schools out of business.  That is already showing signs in the affordable schools, making them less affordable.

The final experience I will talk about in this series is the use of technology.  As with U.S. schools, strategic use of technology in education is incredibly low.  To the credit of education technology entrepreneurs in India, they have convinced schools that they need smart boards and other incremental technology successfully.  However, the actual student-focused use of technology has been minimal to date.

Overall, my trip to India was mind-blowing both in terms of the size of the problem, but also in terms of the energy and urgency which Indians feel to create a great education system.  It was refreshing for someone who spends most of their time in the U.S. being attacked for feeling too much urgency!

My Trip to India - Part 5 - Technology

India is in a very interesting place in a number of respects on technology.  One of the most interesting is the low-cost tablets being created and primarily focused on education.  These tablets will be $20 (subsidized by government) to schools and the people that had played with the Aakash 2 were very positive on it.  This puts India on the leading edge of truly affordable school devices.  Additionally, India has some very large potential advantages on the connectivity side.  They are rolling out a national  fibre network with connectivity even into rural areas.  Finally, they are moving aggressively on 4G LTE connectivity, which should make a lot of sense in the dense urban areas in particular.

So on the hardware and connectivity side, lots going on.  On the other hand, only 6.5% of phones in India are smart and many of those are the lowest end of web access phones.  So it could take several years before the average family has home access.  But overall, if the tablets work, and if connectivity gets deployed, Indian schools could be in very good shape to benefit from online learning.

In talking to the schools, especially affordable schools, technology in a school is seen as a key differentiator, so if there is useful stuff, they could be early adopters.  So far, this has mostly been things like Smartboards.  There is a very interesting pricing structure that has been used here.  The manufacturers lease the equipment to the schools which charge parents for the use, adding a couple more dollars per month to fees.  Maybe the same things happens with tablets, or maybe they are so cheap that parents just buy them.

Anyway, that leaves us with the normal school problem that the online curricula is not great.  Additionally, the majority of Indians other than the upper-class speak a local language and Hindi, but not English.  So even good curricula that is not Hindi localized will have a tough time gaining traction.  Khan Academy for example was localized poorly and I heard plenty of criticism about the translations, though there were clearly still people trying to use it.

My conclusion is that the strong desire for learning in India, the desire to use technology, the competitiveness of the affordable schools, the enormous leverage needed to allow teachers to teach 400 students each, and the hardware/connectivity plans could make India a big early adopter.

Friday, November 2, 2012

Tips for EdTech Entrepreneurs - Navigating This Mess

For many edtech entrepreneurs, their first angel or venture pitches are pretty nerve-wracking.  Here's a chart I will put up to characterize the thoughts of investors who actually want to make money in the K12 space.

The graph above is how I have always thought about startups.  You start in the lower left corner and you have to find your way to the upper right.  During that time, you are plotting a course between focusing on customer growth and revenue growth.  Traditionally, if you are an enterprise company (selling to businesses), you would try to go to the upper left and get revenue traction, especially with some well-known customers.  My company, NetGravity followed this model, locking down media companies for our adserving product early on the Internet and just churning away on sales to $40M in four years.  Situations like the beginning of the Internet don't happen often, so this traditional approach is not always the best.

Modern SaaS companies like salesforce and others changed this path by figuring out freemium models that got them significant user traction.  They give their product away for free with limited features or user caps and get users addicted to convert them to the paid version.  Their monetization depends on the number of free users that they could convert to paid users.  For salesforce, I remember that they were in the 7% conversion range.  Dropbox and other modern Saas companies have also figured out how to be in that magic 5%+ conversion range.  So 95% of your customers use a set of your features for free and you make all of your money on the final 5%.  In the consumer space, this is even more true.  If you figure out a monetization scheme that scales up with customer passion/usage, you will see "whales", customers that outspend the average customer by 10-100x.  Game companies like Zynga are quite dependent on whales.

So how does any of this apply to the way you think about the education market?  The main way it should affect you is the big red X from the upper left to upper right corner.  I have probably talked to 100 edtech entrepreneurs who are going about this market the traditional way, building revenue traction with a few customers and trying to scale up from there.  Here's the problem.  Everyone who has been investing in this space for a while understands the Valley of Death (see my first post).  Most angels and VCs would like to see an exit in 5-7 years.  Since you literally spend your first five years just trying to get sales traction going from upper left to upper right, this approach does not fit well with what typical investors want to see.  You will be very frustrated walking into the offices of these investors.  More likely than not, the "enterprise" partner in the firm will listen for 5 minutes and then throw his body between you and any other partners in the room to try to end the conversation.  He does that because VCs and experienced angels are very good pattern matchers.  Since edtech has been around for 30 years, it's very likely he or a good friend made the mistake at some point of investing in someone going from upper left to upper right.  It didn't work out well and blew a big hole in the floor, and he wants his partners to run the other way.  This is rational behavior on investors part, because that pattern has repeated itself over and over again for 30 years.

So, don't take it personally, but you need to decide what to do about this situation.  Here are a few ideas:

1 - Do you >really< need to sell to schools and districts?  There are some products that you just do.  It's incredibly hard to imagine selling a student information system to parents or teachers.  Ultimately, that's a backend system that helps schools and districts keep track of their kids, so they are the payers.  If you do need to go down this path (and I don't want to discourage you because, believe me, the space needs much better products here) then you need to get comfortable with a couple of things:
    a - you are almost certain to experience the Valley of Death while you get sales traction.  So build your company extremely lean - $50K monthly burn is good here.  You are going to be a Grinder most likely.
   b - go for subsidies.  Whether it is a federal subsidy or philanthropic subsidy, grab it.   Your job is to get through these five years without dying and you will need every penny you can get.
   c - work hard to find other people that can bring your products into customers - If google or microsoft wants you to be a partner on the launch of their next tablet, invest in that heavily.
   d - don't waste a lot of time talking to angels and VCs who are looking for a typical 5-7 year return.  It will be frustrating for you and them.  There are folks in this space with longer time horizons and lower expected returns.  They are dwarfed by the normal institutional investors, but if you go down this route, you are going to need to find all of the patient folks who will stick with you while you grind.

This by the way is the path that Dreambox went down, so I've experienced three years of it.  It can be done, it just takes incredible grit.  The miraculous thing is that once you get out of the other side of the Valley of Death, your company is as good or better on growth and other metrics than most enterprise companies.  This is because the Valley of Death creates a moat behind you, killing off anyone who would like to compete with you.  For investors thinking about the space, the handful of companies like Dreambox will be very appealing.  But that's not going to help you as an edtech entrepreneur when you are launching.  If you have to go this route, put your head down, survive, and get to the point where sales start to feel like normal every day things.  Probably $2-3M in annual revenue is a good point to start feeling better about your chances.  Note that this is not $2-3M in subsidized revenue.  If someone other than schools and districts are paying you, be thankful for the cash, but don't fool yourself that anyone will remember your name when the subsidy is over.  Investors should be very focused on where the money is actually coming from in companies that brave the Valley of Death.

2 - If teachers and maybe ultimately schools and districts are your customers, can you go the modern SaaS route and give away your product to 95% and collect from 5%?  (By the way, getting to that 5% is brutally tough, most companies who get large customer traction can't crack the monetization conversion formula).  Many companies in the space right now are going down this route.  Personally, if you have to sell to the existing institutions, I would implore you to go this route instead of #1 to get around the Valley of Death at least for a while.

Here's the main problem with this approach.  We haven't really seen anyone crack this yet, with the possible exception of Study Island (Archipelago).  They didn't really go freemium, but at $3-$4 per student, it was cheap enough that the market treated it that way.  There are many companies that people are watching here - EdModo, Engrade, Class Dojo and others.  They have a ton of users, now how do they create a revenue model?  One thing I've noticed here is that the "consumer" partners in a lot of venture firms and consumer-oriented angels get interested in these companies.  They have a lot of the same dynamics of consumer and horizontal SaaS plays.  The caveat here is that most consumer folks are used to much larger potential numbers of free customers to convert ultimately.  Dropbox for example already has 50M customers and will likely go much higher if the bigger tech companies can't figure out how to stop them.  By comparison, the entire universe of U.S. based customers (children) is about 60M.  It's true that you will get schools all over the world using your product, but watch out here because to really get large penetration with enterprise type products, you usually have to put some time and energy into understanding market and customer differences from country to country.  Given this relatively (to pure horizontal consumer companies) small market, the way that you get paid matters a lot more.  So the standard - "we are only worried about user traction and we will worry about monetization later" is naive thinking in my opinion.

Serious investors want to at least understand how many hurdles it would take for you to get to $100M in sales.  That's an enormous win, but they go for that goal with every company, assuming (correctly) that most of their companies won't make it.  If you can't build a likely scenario where that happens, you are going to get a pass.  So let's say you really kill it and get 50% market share and 25M users.  If you can convert 5% (again, that's really killing it), you will have 1.25M paying users.  Now,  to hit that $100M sales threshold, you need to make about $80 per student.  That's about 20x what folks like Study Island were able to charge teachers, and it's more than Principals have in disposable expense dollars (about $10K total usually) so it's safe to say you are only going to make that kind of money selling to districts.  Worse yet, districts don't have an $80 per student line item for edtech, so you are ultimately going to try to pull dollars from somewhere else.  I will explain in a later post why blended schools like Rocketship are a completely different and better kind of customer, but for now you should assume that you are back to the Valley of Death.  It's arguably better to be in the Valley of Death with 25M customers, especially if you are able to hit it quickly with 10% weekly growth like the rockstars of consumer and SaaS world.  But still, you are going to be slogging hard for those conversions.  Again, this is the route I would go if I had to sell to institutions, but both investors and entrepreneurs should understand the spending dynamics working against you as you try to skirt the Valley of Death.

I want to pause for a second to make one comment.  If #1 and #2 are so hard, why do people insist on selling to institutions?  When Jesse James was asked  - "Why do you rob banks?"  he replied "because that's where the money is.".  Likewise, enterprise k12 edtech companies asked "Why do you sell to schools?" answer "because that's where the students are."  Like Jesse James though, many of these companies find themselves in jail.

3 - Sell to parents.  Sorry to hold the punchline for so long, I just see many more companies trying to do #1 or #2.  Personally, the next company I start will go after parents.  As an angel, these are the only companies that really interest me.  A few years ago, with the invention of smart phones and tablets, the world changed completely in education, although the institutions didn't really take notice.  Within 5 years, there will be over 1B people with smartphones.  In most developing countries, the conversion from dumbphones to smartphones will happen at double digit growth rates until everyone has one.  Once you have a smartphone or tablet, you can learn yourself.  Period.  So unless as an edtech entrepreneur, you want to go to your grave fighting for the traditional education institutions to survive this massive disruption, I would suggest you think seriously about bringing the education directly to parents and children.

Like #2, this nut has not been cracked yet enough times that patterns have emerged.  People should be watching companies like Motion Math (and the other 10,000 solid edtech companies trying to figure out path #3) incredibly carefully.  The second caveat to #3 is that if you are starting in the U.S., most parents have been trained that education is free.  We just don't pay for things that help our kids learn.  Luckily, this is really not true in a lot of the rest of the world, especially Asia and developing countries.  Check out Bridge International Academies in Africa if you had any doubt about this - http://www.bridgeinternationalacade  So my guess is that we Americans get over the free thing relatively quickly and the rest of the world doesn't have anything to get over.

One other thing to think about if you are an entrepreneur considering #3.  Consumer tech companies who do well typically focus on tech, not content.  That's because content costs a lot more to build and you have to keep refreshing it.  So think hard about how much content is part of your value equation vs. tech.  The more content there is, the more money you have to raise to get a large user base and figure out how to convert users.  Your burn rate is higher, so it's a lot riskier.

Next time I will talk about why blended schools are different for you diehards that want to go enterprise.

Saturday, October 27, 2012

Health care is already blended

Its funny how life gives you insights. Yeaterday afternoon, my wife had some severe stomach pain, so we went to urgent care at the palo alto medical clinic. During our five hour visit, i was struck by four huge similarities between healthcare and high quality blended schools. First, technology was used in several ways throughout the time we were there. The first and the one my wife was most thankful for were painkillers. True, this is relatively old technology, but it is striking how little the adults can do to help your pain and how immediately the medicine helps. So its better for the patients and better for the staff because they dont have to spend time calming you down. This is alao true for the antiobiotics they finally prescribed. Even after five hours and a lot of pain killers, by the time we got home, my wife was in pain again. So i went to the 24 hour pharmacy and picked up the antiobiotic prescriptions. Within 30 minutes of taking the antiobiotics, she was starting to feel much better. Imagine the amount of time that the staff would have had to spend with my wife if they didnt have this simple technology. This is strikingly similar to what happens when we have the right online learning program for the right gap in a childs knowledge. It just works and they learn what they need almost every time. The amount of remediation time from the staff this saves is incredible, and the outcome for the student is better, they learn it faster and better. We dont have enough of the right online lessons to prescribe yet, but that is improving very quickly.

The second striking similarity is the use of technology for diagnosing what ails you. Our entire five hours at the cljnic was basically a process of collecting data thru blood samples/etc and most importantly through a ct scan. Once these were done, the situation was pretty clear. She had an infection in a part of her intestine. This is so similar at rocketship to our focus on using technology to gather data on exactly what a student knows so that we can identify gaps and change their individualized learning plan to specify what activity is next in their developmentally appropriate sequence. The centrality of the ct scan in my wifes healing process and of the assessment and analytics we do to identify and remediate gaps is a direct parallel and the most important aspect of both health and learning.

The third area of the visit that struck me as incredibly similar is the central importance of the adults in the process. The ed industry has somehow come to think of technology as disempowering, thretening to jobs, etc. using healthcare as a corrollary, it is just the opposite. Because people dont have to spend time on low value things, they can focus on making sure you get the meds and assessments you need to prescribe medicine which will make you feel better. That probably works 80% of the time and the last 20% more significant things like operations or therapy need to happen. But because you have taken care of the 80%, the staff has time to focus on the last 20% which is high value add. Exactly the same is true in a good blended school. Much of the assessment and remediation is just part of the day to day system, saving master teachers the mental space to work on the tough challenges. Within the human caital side of health care, the differentiation of roles is incredibly valuable. The ratio of doctors to patients has got to be 1 to 100. Of the 600 minutes we spent there, we robably only had 5 minutes of the doctors time to determine what assessments needed to be done, prescribe the pain meds, and make the prescription once the assessments were done. But we didnt feel bad about that, because there were plenty of senior and junior nurses making my wife feel comfortable, getting the iv going, wheeling her down to her ct scan, etc. in education, we have this idea that if the teacher doesnt do everything herself, it wont go well. Nothing could be further from the truth in a blended school. Because you have associate teachers and paraprofessionals doing a variety of tasks to make sure students are learning, the master teachers play the role of doctors - triaging problems, looking at data, working on the tough problems, and helping the student to apply their knowledge to critical thinking problems (getting even healthier).

The final observation i had is how completely individualized this sophisticated system allows healthcare to be. You are never in a situation where the doctor says - everyone take this medicine right now or im going to gove you a ct scan because thats what we do for every patient. The opposite problem is true in most of public education - we group students regardless of need and teach to the median, clearly a recipe for failure. In blended schools, we focus on dynamically grouping students based on what they need next developmentally. Individualization clearly works better for both health and learning.

It is striking when you are thrown into a foreign situation like we were yesterday, how much perspective it can give you on what you do every day. It reaffirms for me that we are headed in the right direction in education with blended schools. Now we all need to scale it up.

Wednesday, October 24, 2012

Tips for Edtech Entrepreneurs – Part 2 – Looking at Success Past and Present

If you are still reading this blog, you have shown the mostimportant characteristic of an edtech entrepreneur – persistence!  One way to start to think about how to createa company in this space is to look at what has worked before.  As I mentioned, the vast majority of edtechcompanies die in the valley of death during their first five years, whereinitial traction just takes too long. Others make it past that (often in more than five years) but ultimatelydon’t break through to large revenues. There are a few past companies that have made it through that I thinkare important as examples.  I would dividethese into three categories:

Grind It Out – These are typically the bootstrappedcompanies, they survive the valley of death because they just aren’t burningmuch money.  A lot of husband and wifeteams and others.  Probably the bestexample here is Renaissance Learning which sold the popular product AcceleratedReader (AR).  Renaissance basicallyfigured out the school-level sale.  Theypriced AR at $1500, a price at which just a little lobbying from teachers wouldmake a principal buy it to get rid of them. Great strategy, well executed, got to over $100M.  The challenge here is that getting this typeof traction takes a decade and requires you to bootstrap the entire time,because investors just aren’t very interested in these businesses until theybecome cash cows.

Subsidized Growth – A number of companies in the space havebenefited from federal subsidies during their Valley of Death years.  These include companies like WirelessGeneration and Edusoft, benefitting from the passageof No Child Left Behind andReading First subsidies to create assessments for schools, districts andstates.  This is a very effective way offunding through the valley of death, because buyers are relativelyprice-insensitive (it’s found money, so why not spend it).  The challenge with subsidized growth is thatthe money runs out and the old reality sets in. A superintendent that paid $100/student for your product because offederal subsidies may love your product, but since he has a tiny line item forcurriculum or assessment online in his budget, may only be willing to pay$5/student once the subsidy ends.  So youbuild up a company on the subsidy and then have to respond to the crash when the subsidy ends.

Direct to Consumer – A certain number of companies havefigured out how to go around the institutions and direct to folks who buy rationally.  One good example of this is Rosetta Stonewith language learning, though my guess is that a lot of the buyers of thisproduct were older.  Another example isK12, the virtual school operator which has ramped revenue up over $300Mannually by going direct to students. These are very nice businesses, but there sure haven’t been many.

Present Companies


Dreambox is an excellent example of the 10 year overnightsuccess type of company.  It sellsdirectly to schools and districts and had to be pulled out of the Valley ofDeath in its fifth year by philanthropic investors.  Now in its eighth year, it’s executing verywell and likely to join Renaissance and others in the winner’s circle.  But it has been a long journey.

Subsidized Growth

Ed Elements andeSpark are both going down this path with remarkable success.  Ed Elements provides a common dashboardacross multiple curricula for blended schools.  It isinvolved in many districts proposals for Race to the Top District grants fromthe feds.  eSpark is doing the same withits iPad app which helps teachers and schools manage their learning.  As mentioned previously, the lift thatsubsidies bring is a tried and true strategy for getting past the Valley ofDeath, but they will have to figure out how to adjust their pricing for normalschool budgets once the subsidies end.

Direct to Consumer

It’s probably important to define consumer.  The ultimate consumer is the student ofcourse.  Since we are talking about K12,often students don’t have a lot of purchasing power.  So usually direct to consumer in this worldreally means an efficient channel to them through teachers or parents.  I differentiate teachers from theinstitutions they belong to, because they spend $1000 per year on theirclassrooms and have the ability to spend it on your product.  If you weren’t aware of how much teachersspend, take a look at teacherspayteachers, an incredible marketplace largelyfor digital instructional goods for classrooms or donorschoose, which lets outside folks fund the multitude of products that teachers buy. Likewise, parents have always cared about the learning of their childrenand are increasingly taking matters into their own hands.

Three important examples in this category of current companiesare Edmodo, Class Dojo, and Motion Math. Edmodo provides teachers with a facebook-like interface to manage theirclass, make assignments, etc.  They gavethe product away for free to teachers and now have a couple million monthlyusers.  Class Dojo, an application formanaging classroom behavior is an iPhone app which teachers get for free.  They have also gotten to several millionstudents being managed every month. Motion Math sells a series of rich mathematics apps for the iPad.  They sold initially to parents and have sincestarted to follow the iPad into schools and sell whole classroom and schoolsets.  None of these consumer-orientedcompanies have figured out how to monetize their large user bases yet.  On the other hand, they have none of theproblems mentioned in the historical market analysis I gave in my first post.

Next time we'll talk a little bit more about how some of the angels and venture firms are starting to think about analyzing companies in this space.  Of course, paying attention to investors should be only one input into your strategy.  But since the Valley of Death is the bogeyman to edtech companies, understanding investor thinking can be helpful.

Sunday, October 21, 2012

Tips for edtech entrepreneurs part 1 - why is this market so difficult?

A few months ago i vowed that if i was asked the same question by more than 3 edtech entrepreneurs, i would blog about it to save myself time and spread the word faster. Since rocketship is probably the most natural lab in the country for figuring out if new edtech products deliver, i spend a lot of time talking to entrepreneurs. I also sit on the board of dreambox, which has overcome a lot of the challenges in this market. My goal here is to encourage k12 edtech founders to build great products for the market, but to go in with their eyes wide open about the challenges.

I have found that a lot of tech founders are lured by the mission of education and the vast market size. Many folks have hyped this market like it is a big greenfield just waiting for great technology. Unfortunately that is not the case. This first piece is about what makes our market so tough to crack.

Most k12 edtech success is difficult because companies attempt to sell to schools and districts. Principals and superintendents get hundreds of emails and calls about new edtech products (believe me, we do at rocketship). It is important to realize that except for charters, these sales are government sales, which are far different than sales to companies. Charter schools are currently too small a market to build most companies around, so if you are selling to schools, you are selling to the government. Anyone who has done government sales knows that the government acts in far less predictable ways than private companies and usually much more slowly. Unfortunately, startups struggle in all industries requiring government sales because the 18-24 month sales cycle time is often too long for them to survive.

Unfortunately, there are several reasons why as a market, schools and districts are much more difficult than other government sales. First, there is a strong cultural bias against technology in schools. Tech is considered threatening by many. If we automate too much, we will lose jobs. Adding to the tech averse culture, superintendents and principals are relatively senior folks, and may not have used a lot of tech in their lives prior to the last decade. This is not always true, but often.

Adding to the culture problem is the budget problem. Most any industry these days spends a fair amount annually on technology. In education, most of our tech budget has been spent on getting the right hardware and net connectivity for the last decade. The budget for software, especially learning software often does not exist at all, but comes out of a larger category like curriculum (more on the problem with that later). The budget problem is why you see such tiny deal sizes when edtech companies are selling to districts, even large ones. In traditional enterprise software, if you are selling to an organization and are going to have several thousand users there, your deal is very likely to be several million dollars. In education by contrast, you are elated to get a deal for a few hundred thousand dollars and it probably took you a long time to get it.

As if speed, culture and price were not big enough problems, there are three more killers. The first is misaligned incentives. A lot of districts buy software with no concrete expectation for effectiveness. In other words, as long as a vendor has a nice study showing that their product works, the buyer has checked the quality box. The problem with that is that it means if your product is 10x more effective, the market may not even notice. We saw exactly this with dreambox early on. Despite having the only product that adapted to always give a student the right lesson at the right time for their developmental level, dreambox had a very tough time selling the product. At rocketship, we tested hundreds of products and you could literally tell the substantial difference in dreambox within an hour or two. Yet it took a much more vigorous sales effort and a huge infusion of capital to get dreambox moving. Thankfully it is doing incredibly well now, so we are seeing the market reward their product quality.

So if districts arent buying on quality, and we know they are price sensitive, what else causes problems? Here is the killer. If you are selling to schools and districts, your competition is not the other startup down the street, you are competing against some of the most sophisticated sales teams in the world - pearson, mcgraw hill, news, etc. You are trying to take dollars out of their line item - curriculum and assessment - and they arent about to make that easy on you.

And finally, add to the other problems that most buyers would rather see monolithic products/solutions, since it is easier to buy and manage a product from one vendor not many. You will literally see a better product get booted for a mediocre product with coverage of more subjects and grades. That works for the big guys and against you because they have the resources to build large products and you are trying to pick off a specific need.

So just to recap, you have a 24 month sales cycle, anti-tech culture, small budgets and price sensitivity, quality insensitivity, a desire for monolithic solutions, and you are competing against large professional sales forces in a category they already own. That does seem a little insane when you put it all together. Despite these market dynamics, companies have made it through, created significant sales, gone public,etc. However, almost every successful edtech company is a decade long success, and the first five years were brutally difficult. I call this problem the valley of death - the first five years where most k12edtech companies die.

Ok, go out and have a beer. Its better to know what you are getting into before you start, right? And the good news is that there are ways to deal with this market. Part 2 will start to focus on ways to get through the valley of death, maybe even make it shorter or eliminate it altogether.