Espruino Statistics

I was going through some stats for a talk I'm giving, and I thought you might be interested in this...

See the graph below - this shows Espruino IDE use, based on the web server's access logs - so how many times the board description files (which pins/libraries each board supports) have been downloaded from the Espruino website (don't worry, the IDE doesn't transmit any information to me at all - it just asks for the connected board's file).

Brief stats:

• In December, 37000 JSON files were downloaded - so that's the times a board has successfully connected to the IDE, and excluding all the firewalled and cached accesses.
  
• In November, more people were using ESP8266 than every Espruino board put together
  
• Only 1/3 of users had actually used an official Espruino board
  
• There are now 80,000 downloads of binary files per month
  
• 20,000 unique website visitors a month
  
• I get 500 non-spam emails a week
  
• I've written 7000 forum posts to date - that's over 5 a day, every day, for the last 3.5 years
  
• I've made over 1400 GitHub contributions in the last year
  
  

So yeah, big numbers - and it looks like it's really been picking up the last few months!

1 Attachment

In 'Other' are there more Linux/Pi builds or more unspecified boards in general - it is more significant in the last quarter? Great to see the impact Puck.js has made though!

Any info on Espruino board sales, development hours, supplier dealings, build costs, flow, open source hardware, and your personal experience with such that you would be willing to share for helping others interested in getting started with similar projects?

Statistics per country: ie to ensure interest in some distribution channels of Espruino products in other countries to take some support away so that Gordon can concentrate on Espruino ...

Noted micro:bit

Noted gnz probably saw a repost I did :)

Amazing job you're doing, @Gordon! Congrats!

@Ollie I'm not sure - there are a lot of 'clone'-style boards in there with small userbases - it's possible they are DIY things, but Linux/Pi did come relatively high up there if I recall.

@CanyonCasa it's actually hard to get decent sales stats as they fluctuate so much (I only see when distributors place an order). Something like Puck.js, sold at 20-off sells for roughly twice what it costs me to make it and get it into the UK - but that's assuming I'm working for free.

With Puck.js getting released I'm doing ok in Jan/Feb, but before that I was earning about enough to pay myself minimum wage off Espruino sales - luckily the KickStarters help to top that up to a 'normal' salary, but I'd be earning a lot more and working less if I got a normal job.

I work a 50 hour week on average - I'm not sure quite how much of that is dev work, but maybe 30% average - obviously it goes up around a new release and then dies down a little after. The vast majority is just email answering.

Honestly, my advice to anyone looking at doing an Open Source hardware project with a large Open Source software element is: don't. It's a really bad idea.

Others I have spoken to doing similar things appear to have the same issues as me.

OSHW seems to work great as a business - but when you're selling hardware because of your Open Source software you're entirely reliant on people being nice, supporting you, and not just copying your stuff - and that's not a great way to run a business.

Even if people want to license your software, it's very hard to come up with a fair arrangement because any of their competitors could use the software for free...

So yeah, sorry to be the bringer of bad news, but that's my take on it at the moment. If you're just selling hardware and doing a little bit of software, I'd whole-heartedly recommend it though - just keep the value in your hardware, not in the software that runs on it.

Thanks. I don't think you are the "bringer of bad news." It's nice to get some honest answers. I am recently retired and Ok with just contributing but a friend is considering possibly building a business around small project boards.

I do think unfortunately for you your Espruino works carries an unusually high "large Open Source software element" compared to many shield/hat type projects that could potentially sell a reasonable number of boards with much less software overhead and maintenance.

But, kudos to you for your vision of Espruino and delivering on it. For decades I have thought that it would be awesome to have a microprocessor that could run interpretive scripts on the fly rather than fixed compiled C code or overly complicated and insecure OS based hardware. I hope it remains profitable for you to continue. Thanks again.

but I'd be earning a lot more and working less if I got a normal job.

My wife has often said "why can't you just get a normal job?"

For me, running a business is a "random walk" - the trick seems to be to keep at it. This random walk can meander upwards with care and/or effort.

Couple of ideas:

1) Paid support for non-espruino boards - ESP8266/32 in particular. Proliferation will lead to more support calls, so monetise it?

2) Actively engage with, and market to, ESP8266/32 users. It's going that way, the stats suggest it. Your main product is software. So feed the proliferation required for 1) above?

And don't forget to leave "over the top reviews" on tindie to keep Gordon going
https://www.tindie.com/products/gfwilliams/espruino-puckjs/

https://www.tindie.com/stores/gfwilliams/reviews/

1 Attachment

Thanks! I've been thinking about this a lot in the last week - I just gave a talk on something similar at Makercon.de

I think Espruino suffers from this more that others as you say - but I can't help thinking that if there was more of a clear business path there would be a lot more cool software available... I mean, look at how many IoT service companies there are - because there's a very clear path to monetising it (even if it doesn't always work :)

But yeah, as you say it's about the journey. At the end of the day I'm actually managing to pay myself a salary doing my own thing - and something I like - so it can't be all bad.

I think outside of the Silicon Valley VC world where you can just throw money at something and sell the business for millions before it's even making money it's always going to be pretty hard doing stuff yourself.

Thanks for the suggestions - the problem for me with paid support is I don't actually know for sure who is using what board. I guess I could add something so that the Web IDE could 'verify' you as the owner of a proper board, and you could then post on some special part of the forum. Potentially there could be other perks too.

If you didn't have an Espruino board you could get access to that bit of the forum by donating something. Also other board makers could potentially pay to have 'Espruino certified' boards where their users automatically got access to that bit of the forum?

I think that probably hits at the heart of the problem - I think it's awesome to have so many people using Espruino on different boards - that was the intention all along. However, I want to be able to give people who have my boards really good support - but right now it's hard to do that without also spending a lot of time on other people too (which is hard when I have so little time free right now).

Just to add: The other option here is to add a paid service for ESP8266/ESP32 users - where the board makes a secure connection to Espruino.com and you can then program it from anywhere using espruino.com/ide.

I wonder whether anyone would pay for this, but particle seems to be doing ok with that as their business model.

Your average person can't build C binaries and benefits enormously from being able to run interpreted languages where they have prior experience, or which are easy to pick up from scratch if they don't. Otherwise the ESP8266/32 boards are useless to them and they're not able to join in the fun.

ESP8266/ESP32 are capable micro controllers which, given their price, are difficult to complete with. That battle might not be worth fighting.

So, looking at software side, strategically, I think you're in the middle of a land-grab. In the long term, is it better that Espruino is the goto solution for coding on ESP8266/32, or that one of the alternatives like MicroPython, Lua or, ESP8266 Basic, has become it?

One could see more opportunity to earn revenue from initiatives like paid support or premium access channels in the above scenario, than if Espruino was used only by folk who bought an official Espruino board, while the ESP crowd ran something else.

Of course, many won't subscribe, and will get more value than they give - as is the case now - but with scale/volume it doesn't matter as much, or feel quite as personal :)

I don't think you even need to understand how to monetise in the short term (providing you can support yourself) if you get the distribution and Espruino becomes a/the standard, these things should fall into place.

If nobody is using Espruino, but the owners of official boards, you don't have quite the same opportunity, if any.

Having any sort of paid service at all might result in a sizeable revenue stream. I'm sure there are plenty of people on here that feel like they get a lot more value out of espruino boards and software than the cost of buying a board or two. Being able to sign up to a service that you don't 100% need, but is cool and makes you feel you're giving back can work.

Maybe something like a basic/silver/whatever membership could give you a vote on newest features, early access, forum badge and all that. Then a gold/premium membership might get you priority support, automatic custom builds, fancier forum badge.

I don't need a cool forum badge, but I'd sign up for one to support Espruino because it's awesome...

Sorry I'm late to reply to this.

I know what you mean about the land-grab. If I were VC funded I think that'd be a really obvious choice, but it's difficult. That was Codebender's attitude, but they're shutting down now. I think there's got to be some glimmer of a way of making money at some point in the future before I really go for it - but obviously I don't want to put ESP8266 users off.

I'm actually looking at doing a Patreon page at the moment - hopefully I'll be able to add forum badges (I'll experiment with doing it for donations), and will try and expose specific things to supporters through that as well (like possibly the online ide service for ESP8266, or maybe an automatic test system).

I think realistically I need to bite the bullet and start adding proper services to Espruino.com (just using Auth0 like is used for the forum). At that point I can do some much more interesting stuff - like allowing multiple files in the IDE, easy personalised modules, remote login, etc - and can limit access to features unless you're using an Espruino board or are on some kind of simple subscription.

Hi Gordon,

Have you used Espruino with LoRa?

"Semtech's SX1276 chip is a revolutionary new chip enabling wireless communications at distances up to 15km" https://www.espruino.com/SX127x

I did get the SX1276 to transmit/receive packets, but that's it.

However https://www.espruino.com/RN2483 is a module that handles LoRaWAN. I've had that transmitting packets from one module to another (without LoRaWAN) really easily, and am in the process of doing a project that uses it over a LoRaWAN access point at the moment.

If this is a question about potential business then I did consider it, but until LoRaWAN rolls out in to more cities I'm not sure the volume is really there (and it's trivial to stick a RN2483 on an Espruino Pico - @DrAzzy makes a board that connects straight in.

Having looked into it a bit now, rolling your own LoRaWAN network is still pretty rough around the edges. I mean, it's possible, but judging by how little documentation there is, I doubt there are many more than 100 or so people max that have actually set one up and the only a small proportion of those would see the value of running JS.