Amplemeter Updates

Ample Development Report #2

The time has come for Ample’s second community update. In contrast with the initial development period, where we coordinated multiple teams of contractors to build the foundation of our system, this current period has been focused on fixing all their bugs! and finding the right person to join the team full-time. Apart from this, we were accepted into an incubator where we have been receiving a crash course in running a startup, in particular running the startup in a Lean way. This has led to some changes in strategy and we are much more focused now on customer and business development rather than technical development. What’s the point in building all those features if noone wants them, right? On this point, if you are working in a startup I highly recommend reading the Lean Startup by Eric Ries. You are likely killing your company by not doing it.

Customer Development

It’s is our current goal to engage in 3-4 months of customer-centered development (a.k.a. Customer Development) This, of course, necessitates having customers. We are aiming to release two basic forms of the app-meter system and start selling to early adopters in the next six weeks.

Business Version

After speaking with numerous investors and local businesses, we have noticed that there is a demand for energy management systems in the city. On top of this, our technology stack is much more innovative and cost-effective than any of the competition we have seen. We were approached completely unsolicited by a local restauranteur who owns multiple premises and also by a local hospital asking if we could help them reduce energy consumption in their businesses. Installing in businesses gives us the credibility we need to tackle the larger goals and also helps us build the relationships that we will need get our tokens accepted in the city. Since it requires only minor adjustments to our current setup, we are making the adjustments and will start selling to a few early adopters as soon as possible.

Consumer Version

The early consumer version of the app will show only aggregate energy consumption. We will push new features in regular, small batches, observe key metrics and improve each time. The blockchain cap and trade system is one of these features, but not the primary one. Giving meaningful incentives is dependent on identifying key behaviors which is dependent on having quality data. So our priorities are in this order: data > behaviors > incentives. As the lower dependencies mature we can move up the order.

Personnel Development

Working with contractors made sense for the initial push, to get all the foundation blocks in place quickly. Going into customer development requires a full-time and talented team. It’s the worst time of year for hiring in China because bonus season is right around the corner. In spite of this, we have finally added a Full Stack IOT Developer to our team just this week. Philip is the first person in our company to receive a full salary and I am extremely excited to have found and persuaded him to join our team. He has a history of working with startups and has experience working on everything from designing physical circuits to developing for android and IOS. He’s perfect.

Business Development

We have been more focused in communities of impact investors and environmental activists as these are where we expect to find our early adopters and strategic partners. The goal at this point is improving relationships and maintaining a presence in the local communities. This looks like informal business meetings and coffee rather than community events or evening meetups.

Technical Development

While the period summarized by the first update was an extremely positive and productive period, the past few months have been spent in the trenches, repaying technical debt and dealing with contractors not keeping up with their promises. I don’t mind saying it because this is the reality for all startups, whether they say it or don’t. Sometimes things don’t go great.

In any case, we’re through it now. We have completed all necessary rework and the meters are very nearly integrated with the app along with some pretty cool UX features that speed up the setup process. There have been some issues with the data quality (part of the technical debt) which has held up the data collection for training the machine learning algo’s but we hope that we’re down to the last few bugs on that side.

In general, we do not want to develop anything new until we have the ability to push it straight to a customer and get their feedback on it. To be clear, it’s not because we are short of money (we are not), we’re just working smarter. With Philip joining the team, we should get everything tidied up and in customers’ hands within the next few weeks. The technical development section of the next community update should be much more interesting.

Conclusion

This concludes Ample’s second development update. Feel free to reach out to us if you want to know more, talk business or to get your hands on some of our limited first batch of smarter smart meters.

Japanese (From Google translate)

7 Likes

Ample Update #3 - Coronavirus Edition

Well, time flies and we are almost a month later than expected on this update. Despite being near ground zero for the Coronavirus, we have experienced only minor disruption. We lost a month of time. The past few months have been spent scaling up our production and systems as well as creating larger and more extended installations. During this time I have also completed the Startup Leadership Program where I learned a heck of a lot about being an early-stage company as well as gained some valuable connections and credibility.

Customer Development

Public Pilot

We created an open call for volunteers join our public pilot and over 20 families signed up just from the wechat reshares. We have begun to install in these homes. Having these users will enable us to test our behavioral incentives in the real world, and this is where we will be testing our tokenized energy consumption rights. It’s crucial and we are excited to finally be getting on with it.

Business Pilot

About a month ago, we did our initial installation at our first business location. It is a coworking space in Shanghai. We have identified potential energy bill savings there of over 50%, and this is without making any major changes. Just some behavioral nudges and a little bit of automation. We are working on some automation devices now to meet this need.

University Pilot

The university pilot is on hold for now. The university has switched to distance learning because of the virus. The lessons we are learning in the coworking space and the system upgrades we have undergone to make it work there will be directly transferrable to the university, so we will hit the ground running there when it reopens.

Technical Development

Hardware

Just after the last update, we created a second version of our smart meter. This one is way more accurate than the original (over 99% accurate) and more importantly, we have modified it so that it can be built by assembling only third party components. This means that we can go from ordering parts to installing 200 meters in about two weeks flat. Gotta love being in China. Initially, we upgraded the version1 to include the more sensitive digital converter. Today we received the first delivery of brand new v2s.

Once the devices are received, they undergo rigorous testing to pick out faulty devices before they get lost in an installation. Here is the testing after upgrading the gen1 boards.

One day when I was out, the twins joined in the testing

As of last week, we had installed about 50 devices across two businesses and six homes. We are getting increasingly more organized and improving the quality and flexibility of our installations all the time. This is getting ready to go and install at the coworking space.

Software

Likewise with the software, we have implemented more rigorous testing to catch breaking changes before deploying them in the field. The firmware for the devices was “finished” recently. It is stable enough now and we will probably not mess with it anymore at this stage in our development. The full focus now is on the app and incentives. We have a new app concept that we are working on. We have reused a great deal of our original backend but will be giving the frontend a complete makeover. I will share more on this after it has been verified with user tests. We have also carried out initial machine learning experiments on our data to validate our hypotheses on how this would be carried out and they went well. We wait now for more data and to have a platform available to prove directly whether it is useful or not before committing more resources to it.

That’s it for now. Appreciate you reading until the end and your continued support. As usual, if you are in China, especially Shanghai and want to get your hands on a meter and join the pilot, send me a DM. Until the next update!

Regards,
Andrew Hoban

7 Likes

Very good job.

1 Like

Great update - awesome progress on the hardware, and congrats on the pilot!

3 Likes

Hey guys. You can see some of our new screens and a nice big Symbol plug on our updated website here: www.amplemeter.com

Now if only we had Symbol…

We’ve begun testing now on @crackTheCode’s devnet . Really recommend anyone to use it if they want to do some quick testing. The more the merrier! And thanks Bader for providing it.

6 Likes

Ample Update #4 - A Year On

It’s been over a year since we closed our Angel Round and I am very happy to announce our progress to you all now.

Technology Overview

  • The Blockchain Cap-and-Trade system is implemented and in alpha testing. We will release it to our pilot next Friday.
  • Our load disaggregation algorithms so far have a precision of 75-99% in alpha. testing. If you don’t know what load disaggregation is, you can read more here.
  • Hardware meters are running well apart from one or two final, non-critical bugs
  • CI/CD implemented for all components. App updates releasing once per week.
  • Ample Panels and Sensor Kits complete and stable.
  • Automation devices are on hold for now, no time.

The Ample Panel

The Ample Panel we created in response to a need we saw when running our early pilots in public spaces. It has two purposes. The first is to check and communicate that the environment is safe to work in and the second is to reduce behavior-based energy waste. There are being deployed alongside our meters in two commercial establishments in Shanghai and Hangzhou. We use smiley faces to set temperature anchor points and the panels have been successful in reducing consumption by a significant amount, still to be measured formally but around 20%.

Ample App

The results from the first version of the app have been mixed. While one home changed their habits and reduced their consumption by almost 50%, others showed little change. The weather has been quite varied and we still need more data on this. Our first version of the “Habits” tasks (where users follow some instructions to measure and modify energy consumption in their home) I believe has been a failure. Perhaps when we add the cap-and-trade component we can improve interaction around this but I believe we need to redesign it to be more visceral and gamified and to include output from our load disaggregation. Having such failures is exactly the point of doing user testing and so I am not overly concerned. We will redesign and test again, that is the name of the game.

You can see some of the screens of our app on the website. www.amplemeter.com

Business Development

We have been reaching out to large energy companies with intent to obtain a larger scale paid pilot for our technology stack. We entered a competition for EDP, the top 5 wind producer in the world and one of the largest energy suppliers in Brazil and Spain. Of 820 applicants, we made it to the second last round, so the top 7% but we did not make the final cut to the last 30 who all got to plan pilots. The feedback we got from them was good and I believe we would get in easily one we have a bit more evidence to back up our claims.

Apart from that, we have been scouted by a fortune 500 top ten company and will be meeting with them over the coming weeks to discuss the possibility of a pilot. We have a few other similar leads which we are working on one-by-one.

In the past couple of weeks we won two awards. One was a “High-Level Talent Grant” in a city just outside Shanghai. I gave the pitch in Mandarin (my first Chinese pitch) and we were awarded a small grant. We are now in discussions with the local government and university about running a pilot in the dormitories on campus. The other was with a local accelerator who we are now entering a partnership with.

This section could go on, but I think you get the idea.

Team

Due to COVID, our whole team has taken a significantly reduced (from an already low) salary to extend our runway and reduce our exposure to the global uncertainty in any way possible. Despite this, we still managed to onboard a talented machine learning expert and our machine learning project is finally getting some legs. We have runway and can afford to pay wages well into next year but will begin fundraising soon so that we can finance the larger pilots that are in the pipeline.

Linkedin

Please follow me and ample on linkedin. I am posting there once a week or more now so you can get more regular updates through that channel. Thanks for your support everyone. As always if you are in Shanghai, contact me and join our pilot. You will earn real money!

Kind Regards,
Andrew Hoban

7 Likes