A few thoughts on early stage IOT startups in APAC

After seeing 500+ IOT themed companies in the last 2 years in Asia (ex-China), a few observations.

1. Consumer hardware is a very small subset of IOT

We have seen simplistic definitions of IOT = Wearables + home automation, ignoring a wide swath of other industries impacted by sensors and connectivity. Most attempts in consumer hardware are not so graciously called gadgets or toys. We at SeedPlus, have decided to stay clear of consumer hardware. We like companies that haveout of the box ambitions to build a vertical-specific software stack, an OS + application layer.

2. Expectations of retailers and price to place a device on the shelf has dramatically increased

Consumer hardware companies cannot stop at one device any longer. The bar has gone up. Retailers expect device makers to have a roadmap of device launches, with a commitment to stock the first 2 devices. Opportunity cost is high for shelf space — some other larger device company is not getting that spot on the retailer’s shelf when it is offered to a startup. The demand for a roadmap is perfectly rational for retailers. They have seen many startups die after product # 1. However startups can barely get their first device out. Most run out of $ before the second device is even planned. That really leaves startups with non-attractive options of own e-commerce store or Amazon. Selling goods online is a completely different animal from building beautiful hardware products. Few startups have tried enterprise or telco partnerships with limited success.

3. Make hardware an extension of the software or service:

All bike sharing companies use some form of hardware on the bike to track these vehicles. They did not start with building a great bike but a point to point transport utility for a small fee. If they had decided to build bikes, perhaps the outcome would have been very different, with the possibility of limited scalability. Most IOT companies we have seen are focused predominantly on building great hardware. The hope is, once enough (read: millions) of such devices are bought by consumers, companies can easily turn on a subscription service or monetize data from these devices. Unfortunately, we have not seen it work consistently enough for us to know if the next startup using this strategy is a mega-hit or likely to not hit scale. Instead, we prefer a service or software that solves a specific problem using an optional component of commoditized hardware to enhance user experience. An example would be all companies trying to build OBD2 based telematics solutions. Most focus on the build of the device instead of creating a new version of AAA roadside assistance service where the device can be offered for free for an enhanced user experience or better quality of service.

4. Most APAC IOT founding teams do not scale/ transition well :

Most teams start being hardware builders, unable to make the transition to application or software. Getting a good set of advisors who can help with building a commoditized hardware device is better than being stuck with amazing hardware without knowing how to build a software or data business on top of the hardware. This also applies to hardware companies trying to build their own storefront to sell on their own store. Most land up being less than average. Having an experienced e-commerce operator in the founding team can help.

5. The cost argument does not really work. We prefer ideas that protect against unlimited downside

Most companies in the IOT b2b space like to target perceived low hanging fruit like energy efficiency. If done well, energy saved can be lead to enterprise customers paying for the product. However, what really matters is where energy saving lies in the top 2–3 priorities of any company. No CFO wakes up in the morning thinking she will save 1% cost by using questionable new technology. She has other levers to achieve those goals (forex management, lesser materials wastage, manpower reduction). We prefer companies to build solutions that offer a pseudo-insurance policy by collecting data via sensors, to expose and inform unlimited downside risks. An example could be preventing incidents via early warning data from sensors or devices in retail stores. Our portfolio company Qopper does just that. If detected or predicted accurately, most companies will pay irrational $ for this data.

6. Companies should exhaust the capabilities of the mobile device before installing external sensors.

This may sound counterintuitive coming from someone investing in IOT, but mobile devices have incredibly powerful sensors, available at scale. Done right, the data from a collection of mobile devices can exceed most good IOT sensor networks. Most companies working on a problem need to figure out a way to completely exhaust the capabilities of mobile devices. Often, teams are fascinated by building external sensors whereas mobile devices can deliver acceptable solutions. A good example is the 20+ OBD2 dongle makers we have seen in the past year. The best idea in this space is manipulating data produced by sensors on the phone to deliver more intelligent outcomes.