AI at Amazon, Impact of Self-Driving Trucks on Labor Markets, Go-to-Markets & More
|Feb 20||Public post|
Happy SaaStr Week! If you’re in SF, give Hyder a shout (hyder at dynamo dot vc) to enjoy a hot or cold beverage.
Given the hullabaloo around SaaS this week, I thought it made sense to include a few thoughts on SaaS businesses we encounter. Specifically, those businesses that have an IoT component on which there is a software attach. Many investors shy away from a business when there’s hardware involved. For us, the reality is we’re in the business of using bits/bytes to move boxes and that requires translating the physicality of the supply chain into 0s and 1s using hardware. Our portfolio has several examples of this including Odyn and Numadic.
The presence of hardware does require some varied strategy to overcome unique challenges. I particularly work with founders to focus on the below:
Commoditize the hardware layer. The exception is where there’s a defensible advantage and an order-of-magnitude value add of not doing so. When possible, by off the shelf of try to use an existing installed base. If part of the sale requires IoT, prep to “flood the market” at Series A. That being said, I try to spur early efforts around open sourcing designs and building consortiums around such designs.
Cash management. Cash management is no easy task in SaaS as CAC is incurred upfront. Now imagine you need to place volume orders for your sensors which have prepayment requirements longer sales cycles due to a customer base riddled with “behavioral debt,” cash becomes more difficult to manage. From my observations, startups in this realm require larger sums of cash upfront to prevent high-stress situations.
CAC + Cost of Hardware. The common adage is that LTV must be 3.0x CAC. In the same vein, I tend to lump in the cost of hardware and up the ratio to 4.0x (in year one, at least). This might seem a bit odd but in most cases (if not all), the IoT is the harbinger of software value and it has a cost associated with it. Given that VC won’t be around forever to subsidize the cost of hardware, it’s prudent to try to recoup the cost of hardware in 12 months when or where it’s not charged for.
How’s the software generating better value as a result of the IoT? This might seem like a no brainer but the software should be an order-of-magnitude better with the IoT than without. How is that occurring? Is it capturing complexity otherwise difficult to quantify? How about existing data in customer systems? Does the data facilitate intelligence that changes the nature of the customer’s business? Increasingly, we’ve seen IoT provide a data advantage in the realm of ML/AI where a moat is established around training sets.
What’s the ROI? Related to the point above, what’s the ROI to the customer? What’s the metric you’re selling on that a buyer can quantify and track improvement of? At the end of the day, if you’re not moving the needle in a significant way, you’ll struggle. The presence of sensors generally requires more upfront work or the development of a new behavior (ex: tagging each container with a device or installing a device in the OBD port) which requires outsized value to be delivered to users.
Hardware-reliant software is not new and cannot be “written-away” in such a physical industry. That being said, there’s a prudent way to build a venture-backed business when you have silicon in tow.