The new vocationalism

In the past 30 years, “vocationalism” became a dirty word in education, especially among those of us serving low income students. VocEd meant a two-tiered system in high school with a few kids getting ready for college and the rest stuck in shop class, getting ready for something “less.”

But the ROI of a four-year college has changed, and our perception of the relationship between education and work has to change with it.  The reality is that most people don't go to college to stay in academia.  They also don’t go to college to “self-actualize” (although that’s a terrific outcome when it happens - whenever in life it happens).  They go to college to get ready for, and then get, jobs.  Don't get me wrong, for many of us, college was one of the most transformative experiences in our lives.  We built the beginning of our professional networks and gained a little bit of perspective on life.  These non-academic benefits are undeniable.  Yet the cost of college only works financially if you can get a good job afterward. That may or may not mean a traditionally “professional” job, but it should mean one that enables you to keep moving forward for the rest of your life on all your goals, personal and career.

We are moving into an era where people will, thankfully, have much more clarity about what jobs are out there.  They will also know what skills they really need (not mere guesswork) to succeed at them.  In order to match people to the jobs available, and help people fill the gaps in their skills that will keep them from jobs they want, we have to figure out how to evaluate and get people those skills as efficiently as possible. This is a new, middle-class “vocationalism.”  

The best examples of the new vocationalism are the developer academies springing up daily in San Francisco - dev bootcamp, app academy, hack reactor and the like.  They take untrained folks and in a couple of months give them the technical skills they need to start their first job as an engineer.  In many ways, the developer academies were a reaction to a perfect storm: a glut of college grads with debt but no jobs, and a tech sector with a ravenous need for engineers.  Crucially, tech companies can measure the real skills of a candidate in an interview and tryout process, not by guessing from paper credentials by looking at developers code on github and giving them tasks to do.

These dev academies live in the physical world.  They provide intense bonding experiences over a couple of months for the people going through them.  They are also the first signs of the new vocationalism.  Many people that went to these schools skipped college and are now thriving in high paying jobs.  Companies don't have to care that the students skipped college because they have relevant, valuable, measurable skills, and demonstrated they had the resilience to survive this intense process.  I have hired a couple of people for my engineering team with these backgrounds on an otherwise very senior team. They are smart young engineers with humility, energy and passion along with their skills - exactly what startups crave.

Will this new vocationalism in coding move to other sectors?  If employers become clear about the skills they need (not as easy as it sounds, since 70% or more of what experts decide and do is unconscious, or tacit, expertise), and the jobs are plentiful while truly skilled supply is limited, we will see one sector after another get over their fondness for college credentials and focus on what mastery looks like for a particular job. Even now, employers struggle to make newly minted BAs ready for work - in health care, for example, new nurses often require tens of thousands of dollars worth of retraining time before they can be effective.

Will this new vocationalism move online?  Again, it seems at least some of it is very likely to do so. (But not all: the first time a professional draws your blood should probably not be with you, the patient: “Sorry - this worked so well for me on-line!”)   The way Zeal's engine (and I think most other online learning engines) works is that if you can quantify outcomes for which on-line training helps, we can pick a path for you to attain that outcome through a set of learning experiences tailored to your specific existing skills and gaps.  We use data from everyone that has come before you to optimize these paths.  We can pay attention to specific learning characteristics you have by clustering you with others who seem to benefit from similar lessons.  That constant mapping and remapping of the best next experience tailored to you makes technology-enhanced learning's potential incredibly valuable.  

The current set of MOOCs, with their mass-produced, pre-recorded lectures have missed this  point entirely.  But don’t worry, when the goal is learning, and the market spends $4 trillion on it annually, entrepreneurs will bang on the problem passionately until you learn what it takes to make you successful at work.   

Since vocational training is where the jobs are, it will get a lot of focus, and innovation will flourish.  Bror Saxberg and I have been chatting about how Kaplan, the large education company, is working on this challenge- he’s their chief learning officer. Kaplan is beginning to work both internally and with companies on mapping expertise (both conscious and non-conscious) in a variety of fields in a systematic way (using evidence-based techniques like cognitive task analysis), and then backwards mapping those outcomes into courses and training (also evidence-based, to maximize learner success) that Kaplan can offer. This point seems obvious, yet Kaplan is one of the only higher ed organizations to approach learning in such a systematic way.  Udacity's refocus away from college and into this area will look smart in hindsight.  My guess is that their experience with San Jose State helped them to realize that the other 99% aren't the same kinds of students, nor have the same kinds of goals, as the current 1% of the world's kids who go to traditional universities.  It takes much more to help them succeed, to gain a meaningful career (along with other benefits) from their studies.  But it can be done, and it can be done at scale online, building the skills that students need to join the middle class.

So, if challenging domain-specific skills of all kinds move online, what purpose does our current brick-and-mortar school system serve?  Many people feel a lot of angst about this.  Here's the secret that every employer knows: domain-specific, knowledge-intensive skills are not the only skills needed for expert performance.  What are often called “soft skills” within a domain, involving social emotional intelligence and character skills, play an enormous role for success in jobs (and in life).  

When we try a new engineer out for a couple of weeks, we are actually assessing 80% their character and interaction skills at work and 20% their ability to code.  In any human activity, the need for strong personal and interpersonal skills, not just technical skills, will always be critical.   People who can't work with others, organize tasks, control their frustration, communicate, convince, persist, etc. will not do well even when they have the specific job skills.  Beyond the world of pre-kindergarten, we have massively under-rated teaching these kinds of skills in our education system.  

My hope is that schools at all levels - K-12, vocational, and college; online and physical - start to quantify and measure these skills much more precisely so that they weave job-relevant character building and communication into every interaction with students.  KIPP has started to do this and my bet is that it will pay off enormously in their college persistence rates.  This is where the in-person experience shines, when people can work together over long periods of time to model, coach, and change each other’s behavior to fit the character and communication skills that make you happiest and most successful with others.  Reid Hoffman of LinkedIn wrote a great book called "The Startup of You" on a particular form of soft skills (long-term planning and networking) and their importance in navigating modern life.  I think we can all look at our successes in life and realize that they are a lot of non-academic skills that matter.

This jump to the new vocationalism, both from a technical and human standpoint, is not unfolding evenly across all sectors. The perfect storm is moving to our entire country - we have millions out of work, millions of jobs unfilled, and a national need to regain our place as the most innovative and productive country in the world.  The new vocationalism can help to make the transition from old economy to new at the speed we need to make it.

5 things to remember about teachers when you are building your edtech product

1 - teachers are among the busiest people on the planet.  The best thing you can do for a teacher is to save them time so that they can focus more on their students.

One of the biggest uses for zeal is personalizing homework.  It is incredible how valuable saving a couple hours a week of assembly and grading of homework are to most teachers.

2 - a teacher lives and dies by how their school day goes.  Tools that make it go more smoothly (better instruction, better planning, better assessment, better management etc) are what teachers need.

A great example here is class dojo.  Teachers manage their school day with this tool and the adoption has been incredible.

3 - teachers are very social people by nature and care about human relationships.  The more your system unlocks that sociability, the more engagement you will get.

Edmodo figured this out better than anyone else.

4 - don't make the mistake of thinking a teacher's students are your customers.  If you are trying to improve the lives/performance of kids, you should, but remember that if the person choosing your software is a teacher, your value proposition has to be to them.

We spend a lot of time improving learning velocity of students on zeal, but make no mistake that an incisive report or graph that a teacher can use to better understand a student is worth its weight in gold.

5 - if you have not taught, it is a disadvantage, and the only way to overcome that is to spend a lot of time in the classroom user testing your product.

We have more than one teacher on our founding team, but we still spend at least 8 hours per week user testing with teachers and students.

Why Beyond Schools?

A bunch of folks have asked me about the title of my blog.  Since I worked and led schools for a decade, I clearly have a lot of affinity for them.  So why "beyond"?  I think we have reached the limit of what we can load on teachers and schools and expect them to do.  It is a bit like moore's law, every few years the linear improvements that yield to chips doubling in density every 18 months is threatened, and a new solution needs to be developed to get back on the curve.

I think parents in many Western nations went too far in giving up responsibility for our children's life outcomes to schools.  The only way I see our kids making the kind of gains we want is a partnership between teachers and parents.

I taught in very low income schools in Nashville, TN for three years.  I tried to work hard to get families involved in what we were doing in class.  My co-founder at Rocketship, Preston Smith, was much better at this and his ability to bring the community into the school is a huge part of what makes rocketship succeed.  It also made clear how far the norm had shifted at the average American school towards parents passing responsibility to teachers for their kids academic success.  I don't think that is good.  Parents have to pay attention to their kids, because everything doesn't go well socially, emotionally or academically and any of these areas can really take your children off-track.

In very low-income communities, people are often working multiple jobs and have a lot of things requiring their attention just to keep the lights on and family fed.  If education takes too much time or capacity at home, it won't happen.  Often there are language barriers to understand their children's academic needs or parents themselves don't have the academic background to help on their own.

That is a very difficult set of challenges to overcome in order to help your children.  But it has to happen.  Parents have to be clear on what their children need and have to have the tools to help them. The closer we get to every child walking into class academically prepared, the more we shift the norm back to a partnership between parents and teachers.

That is what we are working on at Zeal.  If we can help parents play a bigger role in learning, we think our superhero teachers will have a fighting chance of helping them to grow up to their full potential.

Are schools better than parents?

I am in the nice position as a third time founder that I no longer have to look at the world as a zero sum game.  My success and those of other entrepreneurs in learning technology are correlated in my opinion, not inversely correlated as I remember feeling in my first company.  So one of the things I try to do is spend time with other entrepreneurs, especially first-time entrepreneurs with good ideas.  It was startling to me at netgravity how many truly awful decisions I made as I was learning to run a company, and how much pain and anguish they caused.  A good example of this is that I had brought on two venture capitalists who were both from the client-server world when netg was a client-server company (ie people installed our software on their site instead of paying for it as a service and letting us manage it).  When I saw the first saas companies, double click and ad force, it was obvious to me that the business model would work better and the customer experience would be better if we managed the infrastructure.  I then went through almost 2 years arguing with my board over this and putting together several ill-fated strategies for becoming a saas company.  Because of course my vcs preferred the place they had always made money, client server, and they didnt want netg to be a saas company.  Ultimately I created a skunkworks to build our service which became a much bigger and higher growth revenue stream for us than the client server business and of course my board loved it then.  But those two years sucked because my inability to get us into saas quickly dragged my team into all kinds of time wasters and probably cost us several billion in market cap.

So fast forward to today in learning technology.  Its a traditionally horrible market that is about to become ridiculously good, but only if you respect the drivers that made it suck and the ones that make it beautiful.  The thing that has made learning tech suck for 30 years is that schools are exceptionally bad at buying things.  You can read other of my blog posts for the details, but anyone who has run these companies or served on boards can tell you the ways it sucks for as long as you would like to listen.  The thing that has changed is that students now have access to learning through their phones and other devices.  So you don't need to sell things to schools.  We even have examples of companies like edmodo and class dojo in our space who figured this out and are crushing it.

So why is it that almost every entrepreneur I speak with wants to sell to schools?  They all have their key insights on why their sales strategy will work 10x better than anyone's before.  And of course one things startups do well is to explore a huge number of approaches to a well understood problem.  So these companies fail, usually slowly and painfully as investors learn that in fact the buyer hasn't changed, their incentives haven't changed, and the startups outcomes dont change.

My only conclusion is that entrepreneurs, especially first time ones, are problem seeking and solving machines and the structural concern about whether the customer has any money is secondary.  Not only that, but because most entrepreneurs in this space have absorbed its culture, the idea that your buyer is the parent and not the school is very close to heresy.  I can't tell you the number of very smart people I speak with who cannot accept that schools are not the one and only vehicle to improving academic outcomes.  So the idea of engaging parents seems like it misses the point.  I think both of these biases far outweigh the actual reason most entrepreneurs give for selling to schools, which is that they don't know what parents will pay for.  This last one really doesn't take much imagination or creativity to figure out, so my conclusion is that people have actually never spent any time thinking about it, because of the blinders created by one and two.  I hope for the sake of the kids that we see more entrepreneurs break out of this mental prison.

Minimum Viable Instruction

At Zeal, we are pretty big believers in the lean startup approach.  The fundamental thesis is that you are constantly doing just enough engineering work in order to understand what your customer needs, by doing small experiments and analyzing the results.  You are building the Minimum Viable Product (MVP) to allow you to test whether you understand your customer.  For example, instead of spending several months designing the neatest best game for student learning ever, you introduce small features and watch your user's access patterns to see if that seems to be compelling.  If it is, you add more and analyze again.  This is basically the scientific method applied to startups.   Engineering is a costly task for startups and MVP is designed to not waste engineering time by really understanding what the customer needs before building something.

Anyway, with that frame of mind, it's interesting to think about how teaching and learning works best.  In our current education system, we believe that the teacher is the repository of knowledge and they know what each child needs to learn.  Since assessments are laborious to create, grade and analyze, we do our best with the information we have to address the needs of each child.  But realistically, if you believe as I do that every child is significantly different in terms of their strengths and weaknesses, their learning styles, their social-emotional state all of the time, then your chance of reaching each child every minute of every day is low.  And that really is not the standard anyone is going for in classroom instruction, we kind of accept the inefficiencies because teaching 30 children at once gives us a lot of leverage and we can deal with some inefficiency.

That is quite different online.  Online is much more like a one on one session between the teacher and student continuously.  So if your online system works well, you are able to collect a lot of data on all of the child's needs.  Still, you don't really know what they need to learn without a lot of probing and you don't want to waste time lecturing them on things they already know or aren't ready to learn.  Putting this together with the lean startup approach of Minimum Viable Product, you can see the similarities between engineering and instruction.  This lean startup method is an iterative approach instead of a monolithic approach which we have seen in the past.  So what about instruction?  Instruction is also a heavyweight task.  It takes planning time from teachers and uses a lot of students time when they could be doing something else.

Online, there is really no reason to do heavyweight monolithic instruction.  Instead, you want to pose problems to students and provide minimum viable instruction to them to help them solve those problems.  That switch of the role of problem and instruction is one of the fundamental changes in online learning.  It is only possible because you have unlimited attention and analytic capacity applied to each child.  So you can test what they each need and give them just enough to keep going.

We actually see the concept of Minimum Viable Instruction in other areas, just not that often in the classroom.  Good sports coaches will often focus on a single thing rather than everything you need to know to be good at that sport.  A good golf coach for example focuses on just your swing tempo instead of tempo, positioning, grip, swing motion, etc.  This minimum viable instruction is possible because the coach has analyzed just your game and not a whole class and has decided based on experience the next crucial thing for you to learn.  I think that is the way that online learning will work best.

Naming your company

It was a really interesting process to name our company zeal.  It took several months for a bunch of reasons.  Normally I would think naming a company is just obvious and easy for everyone, but I think it depends on what you want.  I was going to go the route most people do and just string random syllables together to find a domain I could buy for $10/yr.  But in talking to a lot of people in the consumer Internet space, they recommended I spend a little more time and money and come up with a memorable name.  So I did what a lot of people do and tried emailing people who own domain names we wanted.  Maybe 10% even responded when I made hard offers.  So I thought, hmmm, I kind of suck at this, there has to be a different way.  I asked a bunch of friends and they were mostly doing it the way I was.  But I thought, when you buy real estate (which Rocketship did often), you need an agent, so there must be some good agents buying domains.  I asked around a lot and totally struck out.  People were literally just representing themselves.  I finally just resorted to google searches to see if I could find any domain brokers.  Somehow luck smiled on me, because this one guy - Andrew Rosener - had way more mentions and positive reviews than anyone else.  So I emailed him and he was a good guy.

The deal with him was somewhere in the range of 10%-15% of the purchase price in fees to him.  This turned out to be the best money I have ever spent.  I then went out and generated names with the team and other friends and advisors.  We probably generated 150 names.  The first one we all liked was whizkid.  When I sent it to Andrew, he turned it around in a day and told me the asking price (high five figures).  That sounded fine.  So then I went to my attorneys and had them do a copyright and company name search, and the fun began.  Of course, there are a ton of companies and products in the kids gaming and learning space with whizkid or whizkids as names.  So it makes these kind of descriptive names effectively uncopyrightable, and worse, it's very possible that if you enter the space with a name like this, another company can sew you to stop you from using it.  That could come two years into your company when you have a brand established.  So whizkid ended up being a lesson in copyright for me.

Back to the drawing board, we came up with new names that were more general, though generally positive (more the apple, square thing than the descriptive whizkid approach).  From this list of about 50-75 names, we went out to see how many we could get as domains.  There were about a dozen.  We crossed out all but 3 over copyright concerns, negotiated on all 3 and ended up buying zeal.  If we hadn't had a great broker, there's no way that we could have figured out how many of those domains could be acquired in the first place, so would have just settled for something we could get.  Anyway, I'm not saying zeal is the best domain since sliced bread, but a single word, four letter .com with the energy of zeal makes me happy.  Companies aren't made or lost over their names, but having a memorable and spellable one is worth the time in my opinion.  And yes, it cost 6 figures, but we are actually renting it for 24 months on a monthly rent and then spending the 6 figures in 2 years.  We will either be dead by then or that won't be an enormous amount of money for a good domain.

Hoping that any of you going through the same thing can cut a bit of time out of your own process!

Know think do

The way I think about learning is in three parts. First, you have to know a bunch of things to be successful in life. There is no way to fake it if you dont know multi-digit addition. At some point, if you don't, you will put the wrong things into your calculator and give $20 change on an $18 purchase. (This one happened to me last year). So there is a good body of things you have to know. I think this maps cleanly in k-5 where its pretty inarguable that you need to know how to spell and most everything else they teach you. On the other hand, in secondary grades, it is not as clear. Is it crucial for everyone to know biology? Maybe the amount you can look up on Wikipedia, but a whole year? We can debate what things you need to know (which we do periodically, common core being the latest output), but its clear there are things you have to know.

Second, you have to be able to think. You have to use the things you know to solve problems, sometimes in groups, to create new things, to identify patterns. You can be taught how to think. If you aren't sure of this, talk to people who went to engineering school or law school. The methods great schools teach you about how to think are far more valuable (and more difficult to assess) than what they teach you to know.

Finally, to be a successful person, you have to be able to do things. Getting things done has everything to do with your personal and social abilities. Are you persistent? Are you organized? Can you delay gratification and focus on the long term? Do you work well with others? This third category is much harder to measure than the first two but probably accounts for well over 50% of your success in life, especially modern life. Read Reid Hoffman's - the startup of you - sometime to see the product of a very evolved doers mind. Your career and your life is less shaped by the specific things you know (beyond what you learn k-5) and more by how well you adapt to a rapidly changing world. So knowing less and being able to think and do more is going to serve you well.

In this context, we are getting ready for online learning 2.0. Version 1.0 was the simple translation of teaching and learning online by the pioneers like khan, coursera, udacity and others. Pause and rewind is not what online learning 2.0 will look like. What will it look like? Obviously this is what we are building at my new company, so I'm not going to pre-announce anything specific, but I do think there some obvious things. Online learning will teach you almost everything you need to know. Some of that will be through hundreds of hours of practice. Other will be on-demand (arguably for a lot of what you learn in secondary). But anything that requires personalized attention and infinite patience on the other end of the student will ultimately go online. Physical schools, when they are available to a student, will teach you how to do. Physical places and groups will have big advantages in getting students to change their behavior in the way that these personal and social skills require. In the middle I think we will see online learning systems get better and better at teaching you how to think, but this will likely be the domain of institutional learning for a while.

Of course, it is not quite that simple, but its important to acknowledge the relative strengths of teachers and tech so we can think about how they work together rather than the current debate about how tech will replace teachers. Teachers aren't going anywhere, but their roles will change dramatically over the next two decades. They will be much better paid (because the role will be differentiated the same way professors and TAs are differentiated), they will spend much less time teaching rote skills (the knowing, because that will happen online) and much more time teaching children how to think and how to do. I think we will end up with much happier teachers because of this change, but of course change is hard.

So to me this frame of know, think, and do helps you get clearer on how learning will change over the next 20 years. Now that everyone is online (or coming online faster than you can possibly grow your company), online learning companies have enormous market opportunities. I think we will find that the quality of the learning products they build for the 8 billion people out there makes what we have now in edtech feel like the comparison between the VCR and Netflix.

Universities - Get Your Ed Schools Involved Quickly!!!

One of the most striking things to me about the higher education side of online learning is how poor these classes are.  It's really not that incredible that a tiny percentage of students ever finish a class.  Now at this point, university people are saying, "Hey, this is no worse than our classes have ever been!"  Exactly!

The big change here is that universities previously were built on the research reputations of their faculty.  The fact that they actually had to teach was a burden.  Believe me, my wife was a professor for many years.  She actually loved teaching, but she was far from the norm.  Think back to your college experience, how many professors would you give an "A" to?  It was probably strikingly small.  Going through engineering school, I would give an "A" to many more of my TA's than the professors and many more of my peers than my TA's.  So what the universities have put online is literally their worst work - the work of professors who often know nothing about teaching.

I was on a panel a few weeks ago with a couple of MOOCs.  I got tired of hearing about how millions of students were taking these classes.  It's nice to have this much low hanging fruit, but your product can still suck.   So I finally asked my other panelists what they did about students in their classes that were struggling.  The most honest one actually said, "That's the student's problem, our job is to make sure we explain the content clearly."  This insight into what higher ed thinks teaching and learning encompasses is incredibly illustrative.  For the other 99% of students (not the ones the elite universities are used to "teaching"), every student in your class will have large gaps in their knowledge that keeps them from understanding what you are presenting, no matter how clearly you present.  They won't be able to scurry around and fill all of those gaps quickly enough on their own to finish your course.  Filling those gaps is actually what you do for a living.

I strongly encourage everyone in the higher ed space who has a broad mission to educate the students of the world (and not to just educate the ones they already educate, but do it online) to go out into the real world and figure out how to teach the other 99%.  Udacity's work with San Jose State will be the greatest asset they develop for that company if they make it out alive.  It is extremely likely that if they really get it, they will end up completely redesigning the process of teaching and learning in their system.  I really hope others and the universities themselves do the same.

Here is the ironic part of the whole situation.  The education schools at universities, often the ugly stepchild of higher ed, actually have people who spend their time understanding how the other 99% need to learn.  It would be really great for people to walk over to the ed department and ask for help.  I'm not saying ed departments are founts of knowledge, but they at least understand that the type of teaching happening for kids in k-12 gets much closer to what it takes for students to actually learn.

Who's your customer?

One of the things I have been struck by the past few edtech gatherings is that the traditional education technology industry doesn't seem to know a lot of teachers or students.  Coming from the world of teaching, this struck me as pretty bizzare.

But I realized, because the person with the checkbook is the superintendent and to a lesser extent the school principal, these companies know the superintendent in general and only know other people if that's what it takes to make a sale.  The likelihood that they are going to get to know teachers or students well is unlikely (other than their support team on occasion).  That's pretty broken!

It seems to me that one of the only ways we are going to get software into classrooms that actually does things which are useful to students and teachers is if we change the customer.  I don't want to be too simplistic here, but why don't we make the student (and their parent) the customer and the teacher our partner (i.e. the one that let's us acquire customers).  Now the only problem with that is that it completely screws up the revenue streams of the existing companies, because superintendents aren't paying, school leaders aren't paying, and teachers aren't paying.  But really, other than the people running those companies, who cares?  Once students are your customers and teachers are your partners, you are going to build things they like.

Just to make one last argument to the people actually trying to change the outcomes for kids through the software they build (and I'm going to be optimistic here and say that is why most people in this industry build software for education rather than just to make money).  Let's look at this industry for a moment.  It's about a $4 trillion industry globally.  It spends between 1-2% of that on technology.  Anyone who thinks that number is going to get larger needs to learn more about the politics of public education.  So if we do the math, institutions spend $80B or so a year on technology (hardware, software, and services).  That's not too bad.  Unfortunately Pearson, HMH and News are going to get well over half of that.  So everyone else is battling it out for $40B.  Compare that to making students your customer (i.e. focusing on the consumer side of the learning market).  Just to start, tutoring and test prep is $100-$150B per year globally, children's books, apps, etc. are all large markets.  So the consumer side is probably 4-5x larger today (before the shift to more direct learning by students has really happened) with much less friction than the institutional market.

So I will say it again.  Other than the fact that this is not what we do today, why not make the student your customer?

Digital video editing

A friend of a friend told me a great story about being with avid when they brought digital editing to the film industry. The first time he showed the system, the editors said, "you don't understand, if the director comes in and we aren't cutting film and taping it together, he will think we aren't working". It took him two years to convince them that they were experts in composing scenes and not cutting film. Remind you of any other industry? Except 20 years later?

The Shift from Institutional to Individual Learning

As I get my new company going, I've been reflecting a lot on exactly why online learning has been so bad for so long.  I keep coming back to the fact that if I had to start this company today, and sell our online learning system to institutions - schools and universities - it would be a slow and arduous process.  Since the venture community (speaking mostly of traditional high-growth silicon valley venture) understands that education institutions are bad customers, capital would be scarce.  Since the institutions are hard to sell to, I would have to field a pretty significant sales force.  The combination of lack of capital and high sales and market expense just starves the ability to put a lot of money into building great products.

So what has changed?  People are online.  Since curiosity is a basic human condition, it seems pretty clear that when people get online, they are going to want to learn online.  They aren't going to have to trudge down to some building somewhere to find the sage who can teach them.  They are just going to start learning online.  Learning is a ridiculously large business globally, accounting for about $4 trillion of spending each year.  To date, most of that has been going to the institutions, because they were the guardians and providers of knowledge.  In an industry this large, with 2 billion students under 18 and arguably the other 6 billion folks in the world also curious, it doesn't take a very big shift in the way we learn to open up massive opportunities.

Like everything else with the Internet, it seems very likely that the shift in learning online happens long before the spending shift happens.  I think arguably we are already seeing a huge amount of time shift happening, with people learning through Khan, Udacity, etc.  Every time that kind of shift in time has happened from traditional industries to the Internet before, the spending shift happens later.

But here is what is exciting.  That market problem that kept innovation from happening in online learning is going to decrease year by year as learning moves to the individual, and the amount of innovation is going to increase year by year, until individual learning is much more compelling for a variety of things than institutional.  That rebalancing of institutional and individual learning to more accurately reflect their strengths - institution (social, network, collaboration) and individual (personalized, self-directed, student-centered) - is going to create enormous dividends in the number of people globally who get a solid education and the number of people who can go further in their learning sooner because no one is telling them it's age inappropriate.

Kids learn when they are solving problems

It's funny, sometimes I have the same conversation 10 times in a couple of weeks.  Lately the conversation has been about the nature of how kids are going to learn online.  I think the idea that kids are going to learn from pre-recorded video of a teacher discussing a topic is silly.  I have never seen a child at Rocketship learn that way and my friends running secondary schools say the same.  Maybe by late high school or college this changes, but I kind of doubt it.  When you are in a school, I think it becomes very clear when learning happens.  Students who are working on a problem that they can solve learn by trying to solve the problem and receiving prompts and insights from peers or the teacher when they make mistakes.  This eventually helps them get over the hump and be able to solve similar problems with a lot less mental effort.  That's learning.  This happens thousands of times a day in well run classrooms.  For whatever reason, we have really lost this truth in online learning.  First, we don't spend near enough time trying to make sure the student is trying to solve a problem that they have the ability to solve.  Second, when they make a mistake, most of the time a buzzer just goes off and we hit them with a similar problem to frustrate them more.  We lose the teachable moment all of the time.  And finally, when you watch the amount of receptive learning students are expected to do outside of the context of problem solving to "learn" what to do, it often dwarfs the active learning time.  I am really hopeful that the next generation of online learning systems approaches learning from the perspective of the child solving a problem they can solve.