Pro & Teacher hosting plan -- sustainability, next steps

[gabriel (Gabriel D'Amours)]

Hi team,

This thread summarizes a discovery we’ve done on the sustainability of the Pro & Teacher hosting plan, and opens the discussion on what next steps should be taken.

TL;DR – The P&T hosting plan is not sustainable, it has not succeeded as a product, and it has not brought the business growth we hoped it would bring. The most likely outcome is that we’ll abandon it and only keep our Institutional hosting plan. We can discuss it here, and take a final decision in September. There’s also the question of whether we should keep our current P&T clients on a “legacy plan” or have them migrate to another provider.

ticket to log time

As part of a team-wide, general investigation into OpenCraft’s sustainability, I scheduled a discovery to inquire into the sustainability of our Pro & Teacher hosting plan.

The plan was implemented in 2017, with the goal of “offering a turnkey Open edX hosting solution at an affordable price, with dedicated instances, high maintenance, high security and quality support services. We hoped to tap into what we perceived as an important need for robust, dedicated Open edX hosting, and the revenue would help subsidize our development work on OCIM. We hoped to grow this business to tens, if not hundreds of clients.”

The main findings of the discovery are:

1. There is no significant market interest for the Pro & Teacher plan (given the very small subscription numbers over 5 years).

2. We’re making a net deficit of roughly €112 per month for each instance, given that a few subscribers (currently 11, the highest number ever was 14) are sharing a considerable amount of monthly overhead and maintenance.

Since we don’t expect to grow the number of subscribers substantially anytime soon (the cheap hosting niche is dominated by our competitors), our conclusion is that we should probably terminate the Pro & Teacher Plan to focus solely on our Institutional Plan, which is sustainable because of its much higher price.

Review comments on the document and ticket indicate that the move to Grove might lead to savings on the hosting and maintenance front, so we’ll wait until Grove is fully deployed to see how the numbers change.

We’d like to be able to take a final decision around September, and either maintain our current P&T clients on a “legacy” plan, or help them migrate to another provider such as eduNEXT. This is still being discussed, depending on whatever ends up being the cheapest and best solution.

What are your thoughts?

Special thanks to @tikr, @Fox, and @kaustav for the thorough review!

[paulo]

@gabriel

Since we don’t expect to grow the number of subscribers substantially anytime soon (the cheap hosting niche is dominated by our competitors), our conclusion is that we should probably terminate the Pro & Teacher Plan to focus solely on our Institutional Plan, which is sustainable because of its much higher price.

It’s a shame, but it does make business sense if there’s no market for this.

@gabriel

Review comments on the document and ticket indicate that the move to Grove might lead to savings on the hosting and maintenance front, so we’ll wait until Grove is fully deployed to see how the numbers change.

I think maintenance hours are the biggest cost here, right? Even if maintenance goes down, does it make sense to keep the product around given the interest it generates?

Also, do the costs in this account include sandboxes provisioned for development?

@gabriel

We’d like to be able to take a final decision around September, and either maintain our current P&T clients on a “legacy” plan, or help them migrate to another provider such as eduNEXT. This is still being discussed, depending on whatever ends up being the cheapest and best solution.

If moving them to Grove to get lower maintenance costs generates more maintenance overhead, I think we should think about migrating them.

[braden (Braden MacDonald)]

@gabriel I don’t disagree with your conclusions here, but I don’t really understand why our P&T plans have not been successful, and I’d really like to know that.

The highest number of active users we’ve ever seen on a P&T instance is around 40, so no P&T client has ever upgraded to a higher tier.

This seems like the heart of the problem? I guess if people are just using these for random experimental stuff, without institutional backing, they just don’t get any users. And if they’re not getting users, they’re not going to pay for much.

In fact, thinking about it more, if the price is €49+/month, that’s €600/year which is a lot of money. I can’t imagine too many people would be willing to spend their own money on a hobby project at that price. The only way they would is if an institution is paying for it, or they reasonably hope that their instance can earn them more money through selling courses; the latter is unlikely to happen without institutional support either, or at least a great marketing team.

@antoviaque

But it probably makes sense for us to stop trying to offer a “lower” plan than the institutional plan. That would let us focus on what works well for us, as well as help regaining some sustainability, by closing one of our current internal time budget sinkholes. We could maybe even turn it into referrals fee, by redirecting small clients without institutional budget to someone else’s hosting solution, like eduNext’s?

I have a different idea.

My understanding is that the P&T plan doesn’t work because it’s expensive, and because the type of people creating random courses on it don’t have the skills, traffic, and support necessary to turn their hobby / passion project into a successful course.

But what if:

• We can give them hosting for free
• We can let people experiment with Open edX for free
• We can let them create courses for free
• We can help them get traffic and enrollments
• We can do all of that rather passively (cheaply for us)
• We can earn revenue from this
• We can develop a community that gets lots of web traffic, to market/upsell to

My suggestion is that we replace P&T plan with a shared instance like learn.opencraft.com (or Learnaflow.com? I’m not good with names…) where people can create and publish courses for free. We can integrate ecommerce and let people sell their courses, which would give us a 30% cut of course revenues. But people who want to run free courses can do so for free, because that doesn’t cost us much and the traffic it draws only helps make this more successful and helps sell other courses. In order to keep the quality high, and encourage experimentation, we can only list courses once they’re been completed, reviewed, and approved by us.

We would need to seed the platform with some high quality courses - e.g. the open source contributor’s course, the OpenCraft onboarding course, Creative Commons courses like Becoming a More Equitable Educator, etc.

We could even allow white-labelling for a small monthly fee, but probably with the requirement that such sites also list their courses on the main site too.

We could put upselling features right in Studio like “Need more features for your course? Contact us about custom XBlocks…”

I don’t think the platform should sell certificates; the conversion rate is too low and we don’t have the brand to make them meaningful. But if we have high quality content, I do think we can have a mix of courses that are free and paid, where you can only access one chapter of the course until you pay to enroll.

If we had enough courses, we could also use a subscription model where learners can pay a monthly fee for unlimited access, and the revenues are shared with all course creators (free and paid) based on which courses learners actually use. In fact, that’s probably the best way to go here, if we could get enough course content to make it work. Even for learners who just want free courses, the (cheap) subscription could unlock premium features like badges, learning reminders, mobile access (app).

To make it really successful we’d have to do some custom development to add the subscription functionality and probably also a rating and review funcitonality for each course, but I think it could be quite successful if we can find a content niche that isn’t covered by other sites/providers.

Anyhow what do people think about that?

BTW you can think of this as applying “strategic client management” to our P&T instances - how can we offer them much more value and help them grow their business, without incurring major per-instance costs to ourselves? :)

[pooja (Pooja Kulkarni)]

@braden

My suggestion is that we replace P&T plan with a shared instance like learn.opencraft.com (or Learnaflow.com ? I’m not good with names…) where people can create and publish courses for free. We can integrate ecommerce and let people sell their courses, which would give us a 30% cut of course revenues. But people who want to run free courses can do so for free, because that doesn’t cost us much and the traffic it draws only helps make this more successful and helps sell other courses. In order to keep the quality high, and encourage experimentation, we can only list courses once they’re been completed, reviewed, and approved by us.

I think this is a great idea! If most of the P&T clients are using it for experimental stuff, then this should be perfect. We could have a link on the registration page (or perhaps in the footer as well) about the features that could be “unlocked” in the Institutional hosting plan.

@braden

We could put upselling features right in Studio like “Need more features for your course? Contact us about custom XBlocks…”

+1 for this as well.

[kshitij (Kshitij Sobti)]

@braden

In order to keep the quality high, and encourage experimentation, we can only list courses once they’re been completed, reviewed, and approved by us.

I think this is the crux of the reason I am hesitant to go in this direction. How would we review these courses? What criterion do we follow to approve a course?

We don’t have the expertise in this unless we want the site filled with “Getting stated with Python/JavaScript/TypeScript/Ansible” etc type courses. How do we know that the content is accurate if its subject-matter is not familiar to us? How do we know that it isn’t spreading mis-information?

We would need to ideally hire experts outside of OpenCraft for reviews, and that too will come at a cost. I think with all that included, we should be sure that we’ll be better off from a sustainability perspective than we are now.

@braden

My understanding is that the P&T plan doesn’t work because it’s expensive, and because the type of people creating random courses on it don’t have the skills, traffic, and support necessary to turn their hobby / passion project into a successful course.

I think all these reasons also make them unlikely to make a good quality course that isn’t just a blog post with a quiz. Someone who can afford €49+/month right now won’t be able to improve their quality much if they have an extra €49 a month. Someone who can’t afford that won’t either. We should also be able to provide people resources to help here.

If we have ways to help people get traffic and enrolments, we could do that with existing sites hosted with us, and perhaps even create a meta-site that lists and potentially searches through all our instances (if they agree). We can also with relatively little effort create an SSO system that allows creating a unified OpenCraft account that can be used to log into any participating P&T site we have.

I also feel that there is a lot of competition in this market already. There are a lot of websites out there offering platforms for people to put up their content. Open edX allows for potentially richer courses than many of those sites offer, but such courses also need a lot of effort to craft, which is often only possible under an institution.

---

I don’t oppose the idea in general, but I think this is a pretty big project in itself, and involves a lot of work that OpenCraft as it exists right now can’t take on comfortably. Perhaps we could have a project-specific cell for this, with people with specialised skills? This seems like a multi-year project that may or may not be able to generate any better results than P&T.

[giovannicimolin (Giovanni Cimolin da Silva)]

@gabriel

What are your thoughts?

It’s sad to see that it didn’t catch on - after so much work was done on it.

---

@braden

My suggestion is that we replace P&T plan with a shared instance like learn.opencraft.com (or Learnaflow.com ? I’m not good with names…) where people can create and publish courses for free. We can integrate ecommerce and let people sell their courses, which would give us a 30% cut of course revenues.

I like @braden’s idea but I share the same reservations as @kshitij.
for this being handled as a separate cell.

Starting a dedicated thread to explore other possible revenue sources and compare to see how they fare against each other is a good idea. So far we already have a few projects ongoing:

• Listaflow
• SprintCraft
• Grove

Plus new ideas:

• Curated course marketplace with subscriptions (@braden’s idea)
• Create a white label LMS (@braden’s idea too)
• Host other LMSs (Canvas)
• ??? - I probably missed a few other options here as well.

It would be good to use the same criteria we use to judge the above items to evaluate the Pro & Teacher plan and see how it would fare - just to make sure we’re heading in a good direction.

[Fox (Fox Danger Piacenti)]

@braden

@gabriel I don’t disagree with your conclusions here, but I don’t really understand why our P&T plans have not been successful, and I’d really like to know that.

I don’t precisely have a lot of market analysis data for why Pro & Teacher isn’t successful, but there seems to be an intuitive answer: Because we don’t care about it much.

We don’t sink a lot of time and effort into the Pro & Teacher plan because the service it provides is at odds with how we work as a company-- we’re high-touch, premier contractors for the Open edX platform. If someone wants an hour of support from us, they have to fork over 194 Euros-- which is about 4 months of what we charge for hosting an instance up to 50 users.

The way we would make it successful is going to require incredible amounts of investment to achieve large numbers of low-margin customers. That’s our competitor’s niche, not ours, and I’m confident we don’t want to be competing in that space.

600 Euros is a lot of money for an individual person but not much money for an institution, that is true. However 600 Euros is affordable to any institution – even your standard public high school. Hell, even most institutions could afford the base hosting price of our institutional plans if they wanted.

However even in the case of institutional customers, we still don’t win in the long run if they aren’t coming with large development contracts, or the hosting projects aren’t especially massive. Hosting maintenance is repetitive work we try to automate as much as possible, and it can get boring when doing the things that can’t be automated (yet).

We want customers that want to develop big projects and have budgets for lots of hours. This provides the work style (and funds the lifestyles) we want to live. The Pro & Teacher plan’s market segment is highly competitive and the margins are thin. There’s nothing wrong with providing cheap, at-scale multi-tenant hosting products for the platform. But our competitors are designed to work on such margins and we are not, and we should let them stand out where they excel while focusing on what makes us unique in the Open edX space.

Also I have to echo @kshitij 's concerns about trying to take on this project-- it will be very expensive, it’s not sure to work, we already have several projects we’re trying to get off the ground as alternative revenue sources, and, again, our competitors already have solid offerings in this space. We also still have the ‘Jira replacement problem’ coming around the bend, and if we had to choose between the two of them for a new internal project, I’d choose the Jira replacement.

[antoviaque (Xavier Antoviaque)]

@braden

My suggestion is that we replace P&T plan with a shared instance like learn.opencraft.com (or Learnaflow.com ? I’m not good with names…) where people can create and publish courses for free. We can integrate ecommerce and let people sell their courses, which would give us a 30% cut of course revenues. But people who want to run free courses can do so for free, because that doesn’t cost us much and the traffic it draws only helps make this more successful and helps sell other courses. In order to keep the quality high, and encourage experimentation, we can only list courses once they’re been completed, reviewed, and approved by us.

It’s tempting, though it’s true that we need to be realistic about what we’ll be able to achieve. We don’t have the resources to push for a platform – we know very well from our clients how much effort that can be!

At the same time, we actually already host https://courses.opencraft.com/ – which is useful for several things we are already committed to author, maintain and host ourselves: our onboarding course, the open source masterclass (upstreaming MOOC), and a demo course. We also have an Open edX theme in the works – so in any case we are already actively working on that site, and keep upgrading it.

Hosting others there might be more trouble than it’s worth, but there are likely going to be a few cases where we would like to help a project (we have a couple of those sites we sponsor) - maybe we could offer to them a place on courses.opencraft.com in those cases? We could make it clear from the start that the hosting there is free, but any support is paid hourly.

@kshitij

@braden

In order to keep the quality high, and encourage experimentation, we can only list courses once they’re been completed, reviewed, and approved by us.

I think this is the crux of the reason I am hesitant to go in this direction. How would we review these courses? What criterion do we follow to approve a course?

Since it’s unlikely that we will have many people asking for it, we could probably review requests & projects individually? We could also specifically mention types of projects we would want to help hosting (like about open source, or with course content licensed under a CC license, etc.). And if we did get too many requests, we can close the form to make them whenever we want.

PS: Any objection to make this thread public? It doesn’t contain confidential information about specific clients, or that we have a strong reason to keep private, no?

[antoviaque (Xavier Antoviaque)]

@braden

My suggestion is that we replace P&T plan with a shared instance like learn.opencraft.com (or Learnaflow.com ? I’m not good with names…) where people can create and publish courses for free. We can integrate ecommerce and let people sell their courses, which would give us a 30% cut of course revenues. But people who want to run free courses can do so for free, because that doesn’t cost us much and the traffic it draws only helps make this more successful and helps sell other courses. In order to keep the quality high, and encourage experimentation, we can only list courses once they’re been completed, reviewed, and approved by us.

It’s tempting, though it’s true that we need to be realistic about what we’ll be able to achieve. We don’t have the resources to build, run and market a courses platform like edX.org – we know very well from our clients how much effort that can be!

At the same time, we actually already host https://courses.opencraft.com/ – which is useful for several things we are already committed to author, maintain and host ourselves: our onboarding course, the open source masterclass (upstreaming MOOC), and a demo course. We also have an Open edX theme in the works – so in any case we are already actively working on that site, and keep upgrading it. It does also make sense as a good demo/example of what we & Open edX can do.

Hosting others there might be more trouble than it’s worth, but there are likely going to be a few cases where we would like to help a project (we have a couple of those instances we currently sponsor) - maybe we could offer to them a place on courses.opencraft.com in those cases? We could also make it clear from the start that the hosting there is free, but any support is paid hourly.

@kshitij

@braden

In order to keep the quality high, and encourage experimentation, we can only list courses once they’re been completed, reviewed, and approved by us.

I think this is the crux of the reason I am hesitant to go in this direction. How would we review these courses? What criterion do we follow to approve a course?

Since it’s unlikely that we will have many people asking for it, we could probably review requests & projects individually? We could also specifically mention types of projects we would want to sponsor through hosting (eg courses about open source, or with course content licensed under a CC license, etc.). And if we did get too many requests, we can close the form to make them whenever we want.

PS: Any objection to make this thread public? It doesn’t contain confidential information about specific clients, or that we have a strong reason to keep private, no?

[gabriel (Gabriel D'Amours)]

@antoviaque

Any objection to make this thread public?

No, and done.

[guruprasad (Guruprasad Lakshmi Narayanan)]

@gabriel

TL;DR – The P&T hosting plan is not sustainable, it has not succeeded as a product, and it has not brought the business growth we hoped it would bring. The most likely outcome is that we’ll abandon it and only keep our Institutional hosting plan. We can discuss it here, and take a final decision in September. There’s also the question of whether we should keep our current P&T clients on a “legacy plan” or have them migrate to another provider.

While I am sad to see this end, it has been inevitable for a very long time. We were always looking for the mythical growth and sustainability while bleeding a lot of money over the years. I am sad and relieved at the same time.

[jill (Jillian Vogel)]

@guruprasad

I am sad and relieved at the same time.

My sentiments, exactly.

[Fox (Fox Danger Piacenti)]

Conversation here seems to have stalled and there does seem to be a consensus that discontinuing the Pro & Teacher plan is the way forward.

That leaves us with the question about what to do next:

1. Right now my emails to new potential clients, and our website, all have lots of information on the Pro & Teacher plan. I’d suggest we remove most of the information from the website, but keep the ability to sign up for a demo sandbox. We’d just keep the link on file to send over to prospective clients who haven’t played with Open edX before.
2. We need to decide whether we will keep our existing Pro & Teacher clients or else start transitioning them to another team for hosting. My preference is for the latter, since I think both we and the customers would be better off, but I would like to know other peoples’ thoughts.

As @antoviaque pointed out, we’ll still have the ability to sponsor specific courses/projects we want to help as we come across them, and I think it’s great we can do so. If we find a project like that when running through our Pro & Teacher customers, we could make them such an offer. I don’t know that any of our current Pro & Teacher clients fit the bill, but it’s been a while since I reviewed them all.

[tikr (Tim Krones)]

@Fox

That leaves us with the question about what to do next:

…
2. We need to decide whether we will keep our existing Pro & Teacher clients or else start transitioning them to another team for hosting. My preference is for the latter, since I think both we and the customers would be better off, but I would like to know other peoples’ thoughts.

My preference is for the latter as well. See this comment for details – the chances of getting the P&T plan out of debt seem to be pretty much non-existent (especially if we only keep our existing clients).

CC @gabriel @antoviaque

[antoviaque (Xavier Antoviaque)]

@tikr @Fox @gabriel Yes, here let’s just end the offer, and migrate our clients to another provider (or courses.opencraft.com if we want to sponsor some). To make it happen, I think the only thing we need is someone to take on a small epic to figure out the details and implement the change (both technically and contractually). Who would like to do that?

[gabriel (Gabriel D'Amours)]

@antoviaque @Fox @tikr I don’t mind taking this on. I might require a bit of help for the data migration portion, but that’s about it. I’ll schedule an epic next week and do a discovery.