In our last blogpost I looked at how the recent acquisition of M2M Group by Sierra Wireless indicated that the connectivity business was becoming increasingly commoditised. In that post I described how the revenue multiple is misleading in terms of valuing software companies more highly than hardware. Software companies can secure a valuation (or sale price) of 15-20 times revenue, compared to connectivity or module companies that barely manage to scrape 5x and more typically are little above 1x. It might seem that logic dictates that software is the route to great fortune.
This is not necessarily true. The multiples of revenue for the successful sales are much higher. However, what is ignored are the hundreds of software start-ups that never go anywhere. Part of the reason for low valuations on the hardware side of the house is that there is almost always some residual business left that has some intrinsic value even if it has long ago drifted away from profitability. For software firms, that tends not to be the case.
Survivorship bias is a massive challenge in the technology world. Those small fraction of software players who manage to hit a demand sweet spot, typically at just the time when a moneybags buyer identifies a gap in its portfolio, can command 10x+ multiples on revenue. Anyone looking at the market opportunity in IoT will tend to focus on that too. For module manufacturers, for instance, who can typically only command 1-2x revenue valuations, the idea of diversifying into the software platform space is clearly tempting. However, the chances of hitting the jackpot and coming up with (by build, borrow or buy) a winning formula in the software space are low.
As ever, recommendations to organisations to diversify outside of their core business area, usually at substantial cost, are only worthwhile if that vendor proves against the odds to be successful at it. Part of the challenge is that it’s not entirely unheard of. The real success story was Jasper which pivoted from MVNO to software platform company, but its overnight success took the best part of a decade. The multiple in that instance was almost 20x revenue. On the hardware side, Silver Spring Networks’ acquisition by Itron was based on a revenue multiple that was edging towards 3x. Despite being predominantly a hardware company it had also developed a strong management platform alongside. However, the analogy with latter-stage diversification is not a great one, as SilverSpring built the platform capabilities much more organically alongside the hardware/networking offering.
The other part of the challenge is that there are hardware manufacturers who have achieved very high multiples on revenue. ARM, for instance, was acquired at USD32billion in 2016, around a 25x revenue multiple. While it’s far from exclusively an IoT company, and it’s the designer rather than the manufacturer of hardware, it’s worth considering why such a company can command such a sky-high premium. To a lesser extent this was also true of the purchases of Cypress Semiconductor and Creative Chips which sold for four and five times revenue respectively. The difference is that barriers to entry are high. This artificially inflates the cost compared to the lower barrier elements of the value chain. Making modules and modems is an easier proposition, as is reselling connectivity as done by connectivity service providers such as RACO and Numerex.
The software market has the lowest barriers to entry of any space, but nominally commands the highest multiples on sales. However, what this fails to take into account is the number of software players that have fallen by the wayside in the interim (as noted above). The average multiple is hard to calculate without an effective way of including shuttered start-ups. But safe to say that the 20x+ revenue multiple exits are the exception rather than the rule.
The other variable that needs to be considered is scalability. Software platforms are almost universally scalable, allowing the vendor to sell to every client with almost no incremental cost, other than the direct cost of sales. The manufacture of silicon chips is as close to scalable as the hardware space gets. Module manufacture and reselling of connectivity have been much more localised and less scalable. The least scalable function is that of consulting or systems integration. As a result, valuations of organisations performing such functions will necessarily be lower. However, all of this presupposes that valuation is the ultimate goal. Not every company can be a unicorn.
The final thing to consider is maturity. Any business that has not yet reached a steady-state can not realistically be valued based on a multiple of revenues. This naturally over-inflates the value of fast growing businesses when considered in terms of multiple of revenue; they just haven’t reached a reasonable revenue figure by which to be judged.
Diversifying out of core business and into an adjacent part of the market is an overly simplistic solution to any organisations’ woes, particularly if all it is trying to do is boost its valuation. Diversifying to secure synergies is a far more defensible strategy. Having eyes on the unicorn prize means that most hardware and connectivity vendors will miss the fact that high value software exits are the exception rather than the rule.