Introducing TokenTrack

Last week I launched a new web application, TokenTrack. This site enables you to create a token economy — a way to provide positive reinforcement to motivate people to complete tasks. Using the site, you can create an economy, which is a container for people, tasks and rewards. You can then assign value to each task using a virtual currency called tokens. The tasks and rewards are up to you. You can invite people to your economy by supplying their email address. Everything is private, so participants in one economy can’t see or utilize resources you’ve created in another economy. At the moment, every economy is invitation-only, so no one can access your economy unless you’ve invited them.

This site is an example of a web developer “scratching your own itch.” My daughter likes to do academic and creative projects over the summer, but needs a little structure and responds really well to positive reinforcement. I participated in a few token economies when I was her age and remember really enjoying having the ability to work as hard as I wanted, and to be able to choose my own rewards for my work. Token economies are used in education all the time; if you’re not familiar with them, the Wikipedia article is pretty good.

The site is in beta. It’s free to use for now (at least while it’s in beta). We may charge a little bit for premium features or something at some point, but we’ll cross that bridge when we come to it. There are no major bugs that we know of, but let me know (in the comments or via @jeffreymcmanus on Twitter) if you see any funnies or have any feature requests.

Colophon for nerds: I did the site in Python using Flask. This is my first official site launch using Flask (although I’ve been working on a bigger project for a while now — a rewrite of with a new learning management system). I’m finding Flask to be a very productive way to build web sites, more so than the PHP/CodeIgniter combo I’ve been using for the past few years.

Why We Aren’t Looking for Venture Investors for CodeLesson

On the heels of Kirsten’s excellent post summarizing the state of venture investing for education startups, I had a few thoughts. (OK, I actually had many thoughts.)

In Kirsten’s post, she provides four main reasons why educational technology investments aren’t happening at the same pace as investments for other kinds of startups:

  • The cost of doing a startup is low and continues to decline. I agree with this; however, this factor isn’t unique to education startups. And in some respects education may have higher costs than advertising-driven products, particularly once the business matures (see below for an example). Also, the cost of entrepreneurs’ time remains fixed, which suggests an obvious optimization for entrepreneurs: don’t waste your time by meeting with VCs.

  • Edtech founder teams tend to work on “small ideas.” Sure, although there’s certainly no shortage of small ideas in social or ecommerce, and very often the ideas that seem smallest turn out to be the biggest businesses. I think the “small ideas” notion is really a bias against the fundability of a content-driven business; it misjudges the effort required to develop a minimum viable product versus the amount of time it takes to develop a minimum viable curriculum. The amount of effort it takes to develop a minimum viable product is necessarily limited. The amount of time it takes to develop a minimum viable curriculum is potentially unlimited.

  • The rise of edtech incubators. I actually think it’s too soon to see much of an impact here. It’s true there are certainly a lot of incubators now, but we may be confusing activity with results; I don’t know of any incubated companies that have gone on to raise a Series A (although I’m sure Kirsten will correct me here). Also, the vast majority of edtech incubators seem to be focusing on K-12, which is a much tougher row to hoe from a business and traction perspective. Like most incubators I think that 90% of the effort here will (unfortunately) go nowhere.

  • Investors moving into their child-rearing years and prioritizing education (at least K-12 education) as a result of this. Maybe, sure, although you don’t see investors invest in nursing homes and catheter factories just because they’re starting to get old.

One of the ironies that I noticed when I was taking meetings with investors is that a typical venture investor who lists “education” as one of the industry categories he’ll invest in will actually invest in between zero and one actual education companies. I talked to several investors who wouldn’t invest in more than one education company because of the perception that the two companies would be competitive or otherwise conflict. This situation persists even though investors who fund ecommerce or social companies will add dozens of ecommerce and social companies to their portfolios.

(Related: more than 90% of the VCs I spoke with when I was doing VC meetings reacted with surprise when I told them that education is the #2 industry in the US. Really? You have an MBA from an elite school and you don’t know what the #2 industry is? Hmm.)

The cost issue is an interesting one, although it affects every tech startup. The cost of doing an education startup may be lower than the cost of other kinds of startups, at least in the beginning, but I’m not sure that will continue to be case for the life of the business. At their core, most education startups (including my own, CodeLesson) encapsulate some kind of custom content management system. These aren’t too difficult to build (we didn’t even build our own — to get to market quickly we used an existing open-source learning management system). We are currently in the process of replacing this with our own system (or, to use the Sand Hill Road catch phrase, some “original intellectual property”).

But there are costs, particularly in K-12 and higher education sectors, that other startups don’t have to incur. For example, it’s very challenging to rely upon traditional web-based distribution models (either organic or paid) for K-12 or higher-ed products. Teachers adopt products based mostly on recommendations from other teachers, and students adopt products that are given to them by teachers.

One way for ed-tech companies to get around this is to act like a traditional enterprise software company and hire a national sales team. Sales teams are expensive to start up, which is one reason why the cost of running an education startup could diverge greatly from other kinds of startups.

It may seem like insanity to spin up a national sales team for a company that essentially sells content, but the fact remains that there is one and only one home-run success in venture funding for professional education this year, and it’s, which took a $103 million investment earlier this year. Outside observers of the company credit its simple, subscription-based business model and high-quality video content, but I’m not sure these are actually the main factors in Lynda’s success. Just from looking at Lynda’s penetration into higher ed (which is substantial, and almost certainly did not occur organically), I have no doubt that a large part of Lynda’s traction came because they spent the last decade using feet on the street to sell their product on a site-license basis to hundreds (if not thousands) of universities and businesses across North America.

For us, there are a lot of risks to dealing with unqualified prospective investors. This can be expressed in terms of wasted time and opportunity costs, but also in terms of psychic costs; getting up at 4am to have some dick on a conference call tell me my business can’t possibly scale is not what I’d call life-affirming. CodeLesson has been bootstrapped (and ramen-profitable-ish) since we launched it. We have big plans for expanding the service, adding new subject areas, and growing over time. But it’s my expectation that we won’t pursue any meetings with investors for CodeLesson through the end of this year, at least. If we can’t quickly attract an investment that’s 1) greater than what we can earn bootstrapping or 2) at least as big as investments that comparable companies in our space are attracting, there’s no sensible investment deal to be made, at least from our perspective. It would be insane for us to bring a plastic knife to a bazooka fight; much better to live under a rock for a while until our better-heeled competitors with wayward business models run out of money.

(I should mention that the other reason we won’t talk to investors this year is that the work we’re doing on the site now should make a material difference in our traction and it wouldn’t make sense for us to deal with investors until we can demonstrate that traction. But even then, since we’re a paid service, if we do achieve greater traction, we still may not need investors, because to us, more traction equals more money, and it’s far more appealing to fund business operations from recurring customer revenues rather than one-time investments.)

So, to Kristen’s four reasons for lack of investment in ed-tech, I’d add two more:

  • There simply hasn’t been a Google-esque home run in the educational technology space yet. If lack of past large exits is indeed a deterrent to ed-tech investment as I suspect it is, that means that investors are actually more concerned about potential returns than startup costs. And this seems natural to me, particularly if you’re forced to invest on a 3-5 year timeline. It’s easy for an investor to see why Contextual Advertising Network #945 is worth funding because other contextual advertising networks have done well in the past. But it’s not easy to see how Video-Based Learning Site #945 is going to do since there haven’t been any billion-dollar exits in video-based learning yet. (The word “yet” at the end of the sentence is key.)

  • The perception that education startups can’t get to an exit on a timeline that venture investors are comfortable with (which is to say, 5-10 years). The gold standard here is, once again,, which took 12 years just to go from bootstrap to series A. That means that their current investors will be waiting another 5-10 years (at least) to see a return their $103 million. The number of venture investors who have the patience to bet on this situation is, unfortunately, small, particularly when they feel they could do as well or better in a 3-5 year timeline betting on Contextual Advertising Network #945.

So. We aren’t looking for venture money right now and we won’t be pursuing any anytime soon. We’re more or less content to bootstrap indefinitely. I’ve been coding since I was a teenager and teaching professionals to code for 20 years. It’s my expectation that I’ll continue to do this for another 20 years, with CodeLesson as my vehicle. You can follow us on AngelList if you want, but don’t expect us to post another up-and-to-the-right chart there until next year at the earliest.

In closing I should mention that I’ve used is an example for a number of points I’ve made here. I’m not critical of them at all; I’m a great admirer of their business and their success, and I think they’re worth emulating (particularly the part where they didn’t take an investment until they’d been around for 12 years). I also consider their business to be complimentary, rather than competitive, to that of CodeLesson (since Lynda sells pre-recorded videos while we sell access to expert instructors who interact with students).

How To Cancel Adobe Creative Cloud

tl;dr: 1) Look up your Creative Cloud order number on the “My Orders” page; 2) Click here, answer the annoying questions and scroll to the bottom to chat with an agent to have the agent cancel your account. And have a cup of coffee ready, because this process will still take about 10 minutes, even if you have all your information at hand.

Long version of the story: This month, Adobe decided to stop selling boxed software and to go with an all-subscription model. That’s fine with me, at least in theory: I understand the appeal of a subscription model from the viewpoint of the software vendor (and for certain types of customers). I have actually been on a Creative Cloud subscription for the past 11 months. But I’m not really a heavy Creative Cloud user (I use Fireworks 3-4 times a week, Dreamweaver maybe once a month, and Photoshop a few times a year). Since Adobe is end-of-lifeing Fireworks, I’m not getting enough value to justify a $348/year subscription cost, particularly since Sketch appears to be a good-enough alternative to Fireworks for me.

So, I decided to cancel Creative Cloud. Disappointingly but unsurprisingly, it’s very difficult to cancel Adobe Creative Cloud. First, even finding where you’re supposed to manage your Creative Cloud account on Adobe’s convoluted web site is a challenge. There’s a Creative Cloud Plan Information page that contains a cancellation link, but that link doesn’t actually cancel your plan:

adobe_cancel_dohThis page also unhelpfully doesn’t contain your order number, which Adobe’s customer service needs to kill your subscription. In addition, they make you talk to a human to cancel your plan. But before you can do that, you get to go through their automated Customer Support shark-infested moat (which, unhelpfully, redirects you back to your Creative Cloud Plan Information page, which in turn directs you to contact Adobe Customer Support in an endless loop). Genius.

So after a few go-rounds and another ten minutes wasted, I got into chat with Adobe support. Here’s the transcript of our chat:

info: All representatives are actively assisting other customers. There are 1 customer(s) in line ahead of you. Thank you for your patience.
info: You are now chatting with Saroj.
Saroj: Hello! Welcome to Adobe Customer Service.
Jeffrey: hi there.
Saroj: Hi Jeffrey.
Jeffrey: Can you assist me with cancelling my Creative Cloud subscription?
Saroj: As I understand you would like to cancel your subscription,am I correct?
Jeffrey: That is exactly what I just said.
Saroj: Thank you for the confirmation.
Saroj: Let me see what best I can do for your help.
Saroj: May I have the order number,please?
Jeffrey: I have no idea what the order number is. I purchased this a year ago.
Jeffrey: I'm looking at and it doesn't list the order number anywhere.
Saroj: Login to your Adobe account. Go to My Adobe, then click on My Orders.
Saroj: You will be able to find your order number.
Jeffrey: AD00XXXXXXX
Jeffrey: Really surprised this information isn't on and even more surprised that the "Cancel" link on this page doesn't actually cancel the subscription. It's almost as if Adobe is making it intentionally difficult to cancel.
Saroj: Thank you for the order number.
Saroj: I apologize for the inconvenience caused in this regard.  We will surely take this as a feedback and will work towards to improve on our services.
Jeffrey: Terrific.
Saroj: If you cancel your subscription now,there will be the cancellation charge of US $15.
Saroj: Would you like me to cancel this now?
Jeffrey: Why will there be a cancellation charge?
Saroj: The annual plan you enrolled in offers lower monthly payments and requires a one-year commitment.This plan is ideal for someone with an ongoing need to use Adobe's Creative software. If you decide to end your subscription before the one-year period is over, you no longer qualify for one-year subscription pricing.You will be billed at 50% of your monthly rate for the remaining months in your annual contract."
Jeffrey: Lovely. Let's go ahead and cancel.
Saroj: I understand that you would like to cancel your membership, and I will take care of that for you. However, would you be willing to maintain your membership through your annual commitment & avoid the cancellation fee if I offer you the next month of your membership for free?
Jeffrey: I'd like you to cancel immediately, and if you don't do it in the next 30 seconds, I'll call my bank and have them reverse the charges. Is that clear?
Jeffrey: This has already wasted 15 minutes of my time and it should only have taken four seconds.
Saroj: I was just trying to help you out.
Jeffrey: You're not helping me by wasting my time.
Saroj: Anyway please stay online while I cancel your subscription.
Jeffrey: Hey, what choice do I have. I understand that deleting a row from a database takes a lot of time.
Saroj: Sorry for the wait. Please do stay online.
Saroj: I have successfully cancelled your subscription for order number AD00XXXXXXX.
Saroj: Is there anything else I can help you with? 
Jeffrey: Spectacular. No, there's nothing else, other than making the cancellation process available from the web site without my having to spend 10 minutes searching for it and another 10 minutes dealing with a person to cancel.
Saroj: I apologize if this experience with us has been unsatisfactory.
Saroj: Have a great day!
Saroj: Thank you for contacting Adobe.  We are available 7 days a week, 24 hours a day. Goodbye!
Jeffrey: Adios.

So, from this, it’s clear why they require you to talk to a customer service agent; it’s a sales tactic intended to keep you on their subscription plan. But there’s a cost to this: my time, a resource I can never get back. When a corporation wastes my time needlessly, I remember it for many years (looking at you, United Airlines and Sony). A process that should take about 10 seconds (log in, find the link to cancel your subscription, and click it) takes about 15 minutes. And this delay is, in a sense, punitive: Adobe hopes you will lose interest and back off the cancellation process if the cancellation process is sufficiently painful; making the cancellation process easier costs them money in the form of lost subscriptions.

But for Adobe, there are risks to this strategy as well. In years past, AOL (the original gangster of the shark-infested-moat subscription business model) got itself into serious legal trouble by making it difficult or impossible to cancel the service in a reasonable period of time. I’m certain that Adobe will find itself in the same place if it doesn’t address its process for canceling subscriptions as the company moves to a 100% subscription-based model.

Email Providers with Pay-As-You-Go Pricing

Without wading into the kerfluffle around SendGrid that’s going on this week, here’s a partial list of commercial email service providers that compete with SendGrid. I am specifically focusing on products that provide pay-as-you-go pricing since that’s what we need.

Amazon Simple Email Service lets you send 2,000 messages for free each day if you send mail from an Amazon EC2 instance. Otherwise, pricing is $0.10 per thousand or $0.0001 per message. This makes Amazon one of the most aggressively-priced commercial email service providers around. I found Amazon SES to be challenging to configure and lacking some of the dedicated providers’ features such as a deliverability dashboard. We stopped using SES last year for this reason, but as our volume increases we will consider going back to them.

Postmark charges $1.50 per thousand emails (so, $0.0015 per message). They have a free trial that lets you send 1,000 emails for free. They also provide discounts if you’re sending 500,000 messages or more each month (we’ll get there someday). It looks like they have you pay up front and draw down your account balance like MailChimp does, but they’re a true pay-as-you-go service. Thanks to @torrez for the recommendation.

Mailgun, owned by Rackspace, is not a purely pay-as-you-go service since they have monthly minimums. Their standard account is $1.00 per thousand (so, $0.001 per message) with a $19/month minimum. So this seems like it would would be a good choice if you were planning to send in the ballpark of 15,000 messages or more each month. (They have higher volume plans that reduce the cost per message.) Thanks to @jetsetter for the recommendation.

Vertical Response says they charge $0.01 per email, but it looks like that pricing is variable depending on how much you use (we do email blasts to about 6,000 users which they say would cost $72.00, or $0.012 per email).

Constant Contact does not appear to provide pay-as-you-go pricing. They charge $50 a month for up to 5,000 emails, or $0.01 per email. (They have more expensive monthly plans that bring down the price per message.) Because they don’t offer pay-as-you-go pricing, you’ll almost certainly wind up paying for more capacity than you need.

MailChimp requires that you purchase credits ahead of time to get pay-as-you-go pricing. A $100 prepayment gets you 5,000 credits (that’s $0.02 per email, which makes them one of the more expensive providers).

Campaign Monitor charges $5 per “campaign” plus $0.01 per recipient, but unlike a lot of providers they don’t place restrictions on transactional versus non-transactional (marketing) mails. A transactional email is one that’s sent as a result of a user action (like registering for your site, or in the case of CodeLesson, enrolling in a course). When you’re choosing a provider it’s best to carefully review their terms of service to ensure that what you need is in alignment with what the provider’s terms.

I’m sure there are more providers that provide pay-as-you-go pricing, and I realize that price-per-message is not the only important metric. If you’re using a provider you like, please recommend them in the comments.

Three Thoughts on the New PayPal Developer Site

1. The relaunch of is long overdue and a terrific improvement. PayPal has never had a developer web site worthy of its promise until today. Part of this has to do with how PayPal and its parent company eBay were organized: a dog’s breakfast of products and ancillary initiatives that suffered from a lack of coordination and were often in direct conflict with each other.

It’s clear that a lot of thought went into not only how the site is organized and presented, but how PayPal talks about its products. No longer must you know whether Website Payments Pro or PayPal Website Payments Standard happens to be the correct choice for you before you proceed; the choices are very clear and are arranged in a manner that reflects the user’s goals, not the platform provider’s branding strategy:



Starting from a business goal (“Try Our Shiny New Blortz 2.0!”) rather than a user goal (“Make Money By Accepting Payments!”) is one of the most common errors I see in any kind of product web site; it’s particularly common to see with developer portals, which often seem to be developed by retooled consumer marketers who are operating out of their depth when addressing a developer audience.

2. This site was not developed in isolation. The new site reflects holistic coordination between the product/marketing side of the business and the people who are in charge of engaging with developers (which, at many platform companies, are frequently two distinct sets of people who don’t always coordinate with each other very well).

Presenting information about developer products in this way will certainly sand down the rough edges with regard to getting new developers on board for PayPal, so I’d expect them to see a benefit from the new site fairly quickly.

The new site has a terrific information architecture and a clean look. There’s even a bit of Twitter DNA in there (they’re using the tremendous Twitter Bootstrap UI library).

3. Federated authentication is an interesting and useful addition. PayPal has adopted federated authentication for its developer site, which means you have a button on that logs you in using the same credentials you use on the main site. This has a minor immediate benefit for developers, since you no longer have to maintain two PayPal identities — one account where your money lives and another account to access your developer sandbox. But potentially more importantly, it means that PayPal could transform into an authentication provider of its own at some point. This would give any consumer with a PayPal account the ability to log into a web application using PayPal in the same way that we log in using Twitter, Facebook, or LinkedIn today:


The difference, of course, is that the security regime provided by PayPal would be much greater than other federated authentication providers. As a payment provider, PayPal must adhere to international laws regarding data privacy and security, which would seem to support a higher level of trust for the federated authentication scenario. I’d feel much better sharing my personal information with PayPal than, say, Facebook.

Jeffrey McManus has led developer initiatives for eBay and Yahoo! and has consulted on developer platforms for a number of startups, including Twitter and Twilio. He currently leads Platform Associates, a consultancy that helps online businesses develop and manage platform products, and CodeLesson, which provides instructor-led online training for software developers.

Our SF Sketchfest Schedule

sf_sketchfestReally looking forward to SF Sketchfest, which starts this week. I’ve always loved live comedy but I couldn’t articulate how much I love going to shows until we went to a Groundlings show in LA last year with a sports-watching buddy. “This,” I said, as we exited the theater, “is my live sports.”

Last year the entertainment highlight of the year was the staged reading of Wet Hot American Summer and the Groundlings improv show, where I posed for photos with the excellent Jim Rash, a Groundling who stars as Dean Pelton on Community and won last year’s Oscar for best adapted screenplay (“The Descendents”). I heard he sold his latest film for about a bazillion dollars at Sundance this week, and I can’t wait to see it.

Anyway, this year we’re seeing more Sketchfest shows than ever, starting with Judge John Hodgman tonight.

The other shows we’re going to see include:

  • Futurama Live (Saturday January 26)
  • Party Down with Party Down (Saturday February 9)
  • Theme Park Improv (Saturday February 9)
  • Don’t Let the Comedians Do Story Time (with children’s book author Mo Willems, Sunday February 10)
  • An Afternoon with Burning Love (Sunday February 10)

Let me know if you’re going to any of these shows, we can get drinks or whatnot. (I guess we could get mimosas before the Mo Willems show, since that starts at 11 in the morning.)

Saint Etienne, “I’ve Got Your Music”

Didn’t realize that Saint Etienne had released a new album in 2012 until New Year’s Eve. I think this may be my favorite of the year; it’s definitely one of their strongest albums. This video is utterly adorable and features a neat San Francisco cameo. How many of these records are in your collection?