Week 455

I’m in the office and back from Japan today. Attacking my emails. So I thought I’d share a few things that are on my mind.


Berg’s slightly smaller than when I headed out. We’re focusing more and more on the cloud services piece — you can see that on the new website which started rolling out last week.

New strap: Cloud services for hardware innovators

It continues:

You’ve got the hardware. You’ve built the website. Berg does the rest. We’re the missing link. Message and manage all your connected products with our REST API, using familiar JSON. Add web connectivity to your device with our Devshields. Then move to your preferred wireless solution when you go to production with our client libraries.

We’ve also spun out Little Printer to its own site (and blog) where it has more room to breathe as a product.

So this is all in the spirit of getting our positioning really clear: What does Berg do? Hopefully the website is clearer on that point.

There’s a corollary to that positioning question. What doesn’t Berg do? Well, we’re about helping other hardware innovators get to market, and inspiring development; we won’t be bringing new hardware to market ourselves. We’re happy to lend a hand with design services for connected products, but no more iPad apps or strategy workshops. We’re about the cloud services, not the vertically integrated technology stack all the way from prototyping platform through wireless connectivity and mobile UX.

And that means a tighter team. It’s a smaller room than a couple of weeks ago, and there have been some sad goodbyes. I won’t pretend it’s easy (it ain’t).

But there’s a way we can really help hardware startups and the new product groups inside bigger companies, and this is it.

Fitting in

Another corollary to the positioning question is: What does Berg do in relation to others?

That’s an interesting one. In Old Berg – the consultancy – it was pretty clear: We slotted in and performed a design function usually during R&D or early product development, with enough skills in strategy and tech to link up with both.

That’s not the case now; we provide a platform. So fitting in now means a couple of things. First there’s Berg The Platform. That exists in a technology stack that enables connected products. (More about that below.) Second there’s Berg The Company, and the job of the company is to build the platform, attract attention to the platform, and do whatever work is necessary to help companies take advantage of the platform… whether that’s by doing faster prototyping work because we understand the issues, or by providing the outsourced infrastructure for a connected products line.

The platform itself… well, I remember trying to sketch this space 6-12 months ago, and there was no technology stack for the Internet of Things. Now, well, there’s something emerging.


The verticals are emerging, still with fuzzy edges. This is my working model, as at February 2014.

Wearables and appcessories have appeared as smartphone peripherals, enabled by Bluetooth LE (aka Bluetooth 4), relying on the iPhone or Android phone for connectivity, often single user devices.

Smart cities and connected buildings have their own tech vendors. Agriculture, industrial solutions, and medical all seem to be emerging as “Internet of Things” verticals too. I’m not super knowledgeable about these non-consumer spaces.

Smart home is an interesting one. Characterised by programmable, low latency platforms with commodity products, I would place SmartThings, Revolv, Belkin Wemo, and Philips Hue in this world.

Then there’s consumer appliances, which is still emerging and overlaps a little with smart home. Products for the home and office that tend to have shared ownership, have a persistent connection to the web (remote control is a feature), and where there’s potential for the services to be as important as the product. I would put Withings, probably Nest too. Definitely Little Printer. This is where Berg is, and the vertical I find most exciting.

There’s a lot of room for growth in consumer appliances. The services I see so far aren’t particularly imaginative — with most new products, it’s not clear why the product and associated service are joined at the hip. Recipes on a screen on my fridge, really, is that the best service for that product, or the best place for that service? An example of a well-integrated device is the Withings smart bathroom scales, which they position as a “Smart Body Analyzer”. It makes sense to do health coaching directly from the device that you use to make the measurements.

I see two routes to making order-of-magnitude better products in this vertical:

  • Help manufacturers experiment more easily, because that will lead to better services. At the moment they have to adopt a waterfall model of concept then develop (because embedded software is hard to build and their technology is tightly coupled). With a more agile platform using web APIs – which is what we offer – design and development can be iterative and agile, learning from customers without having to back to the drawing board
  • Help companies that *already* understand services to make connected hardware, because maybe it’s the service that’s the difficult bit. Web companies have brilliant customer-focused services. Look at Google Apps, say, or the tens of thousands of startups in Silicon Valley and London Tech City. These services are hard to build. So maybe – when it makes sense to have a physical embodiment of the service – it makes sense for the web company to extend into hardware. I’d love to have a Instagram digital photo frame, and I think that product would sell really sell. But I can’t imagine a digital photo frame manufacturer building an Instagram equivalent. So our platform – by giving web APIs to connected hardware – lowers the barrier to entry for web companies to start building, from prototype through to production.

The technology stack (“Horizontals”)

I’m still figuring this out, but I have some rough notes as to the technology stack for connected products in the “consumer appliance” vertical (or whatever we end up calling it).

The stack looks something like this:

7. Your UX
6. Your backend web service 
5. Device management + services
4. Cloud message bus
3. Hardware wireless solution
2. Connectivity APIs
1. Your hardware

Electric Imp is vertically integrated across 2, 3, 4 and works for prototyping and production. Spark is 1 (kinda), 2, 3, 4 and a complete solution for prototyping.

Berg was attempting 1, 2, 3, 4, 5 for prototyping, with a move to 2, 4, 5 for production. Too much; too complicated.

So with our new focus, we’ll concentrate the platform on 2, 4, 5. i.e., we focus on the high-level cloud services (message bus and device management), and you bring your own hardware connectivity. For prototyping and production, we’ll have ways to plug into common hardware connectivity solutions (more of that in coming weeks).

Our job is to fit in. To bring our high-level cloud services (easy web APIs, developer and customer services tools, and more) to wherever developers are creating connected products, and to provide those services so developers don’t need to build them from scratch every single time.

You can see the beginnings of this positioning on the Berg Overview page. Again, it’s just a start.

Figuring it out

It’s a new world. A tech startup isn’t like a design consultancy. Particularly, it’s not like a design consultancy whose raison d’etre was to be unlike any other studio out there. Now our jobs are clarity, fitting in, inspiring, and solving real problems really really well.

It’s good to know what our challenges are.

Yet… it’s early days. We’re a young company again. A tiny bit of Old Berg – a splinter, a seed – that is being allowed to grow into a whole new thing.

Probably not coincidentally, I recently found myself on the Wikipedia page for the Toba catastrophe theory.

70,000 years ago, a supervolcano eruption in Sumatra reduced the global human population to as few as 3,000 individuals. The richness of millions of years of evolution distilled to the size of a village. It must have been a heady brew to distill; humanity hasn’t done too bad since.

These population bottlenecks, these reboots, these founder events have the effect of channeling genetic diversity and making a new population which has less diversity but is more distinct… and possibly better adapted to the new world.

So that’s what it feels like. A new world of connected products and a new Berg, coming into being right at the same time.

I’ll try and write these notes a little more often.