I recently had the opportunity to speak with Sonny Garg, who leads the energy unit at Uptake, an industrials-focused predictive analytics company launched by Groupon co-founders, Brad Keywell and Eric Lefkofsky. Uptake’s platform serves a broad range of markets – rail, aviation, mining, energy, construction, agriculture and retail, among others – and recently announced a deal with Berkshire Hathaway Energy for the application of Uptake’s predictive analytics to wind plants at two BHE subsidiaries. I spoke with Garg about both the BHE deal and Uptake’s approach to the energy sector.
While a broad range of companies are active within the predictive analytics and PaaS (platform-as-a-service)/SaaS sectors, Uptake stands out for its highly focused business model. Uptake applies predictive analytics to create value by improving three operational metrics – reliability, productivity, and safety. This approach drives lean software development, seeking to achieve these outcomes through the simplest, fastest, and cheapest way possible. The approach also extends to the revenue model, where Uptake’s revenues are tied to the value its software creates, unlike most other PaaS/SaaS players who price on a subscription or fee basis.
The exclusive focus on predictive analytics differentiates Uptake from many other IoT players. According to a McKinsey Global Institute report, “most IoT data are not used…information is used mostly to detect and control anomalies – not for optimization and prediction, which provide the greatest value.” Even within companies that provide optimization analytics, providing only the analytics differentiates Uptake from competitors such as GE Predix and C3 IoT, who have a “middle-up” model in the pyramid (see right), starting from the enterprise software layer and moving up to the analytics. Garg highlights that, by working from the analytics down, Uptake only brings in the data that has value in delivering the targeted outcome, thereby minimizing integration complexity and time. In his words, Uptake has built an “actionable insights platform” with a single code base that can be upgraded every week over the air, in comparison to enterprise software that has time-consuming, frequent updates.
Uptake’s strategy in the power sector has been deliberate and highly aware of past challenges that have hampered SaaS companies in this sector. Garg, as the former CIO of Exelon, has watched the evolution of SaaS business models in the utility space over the past decade, and the lessons Uptake’s predecessors learned have certainly shaped the company’s customer segmentation and product development approaches. The BHE deal displays several facets of Uptake’s strategy, both across sectors and within power:
- Seeking to work with iconic brands across industries. Caterpillar (who is also an investor) and Progress Rail are two prior examples in other industries.
- Addressing specific, measurable improvements. In this case, the improvements targeted are increased availability and reliability of wind farms. This demonstrates both precision targeting of use cases where value creation is quantifiable and measurable, and an example of the cross-industry synergies that Uptake realizes. In this case, the understanding of vibration that Uptake has developed through prior experience with diesel engines in rail means Uptake has both a broader dataset of relevant data, thereby improving prediction, and a head start in understanding how to improve wind availability, even without prior direct experience in wind.
- Targeting streamlined paths to sharing in value creation. The specific BHE subsidiaries involved are BHE Renewables and MidAmerican Energy Company. BHE Renewables is an independent power producer (IPP), meaning it is not subject to regulated profitability in the same way that utility companies are. This means additional value for BHE Renewables simply means additional profits, which can be shared with Uptake. On the other hand, additional profitability for regulated utilities can actually cause challenges, as utilities are subject to regulated rates of return and profit margins.
This third facet, dividing the power sector into T&D utilities and IPPs, reflects a clear desire to avoid the challenges that have hampered SaaS companies that serve the regulated utility segment of the sector. Utility procurement cycles, characterized by long, RFP-driven processes, have been a major issue for start-ups. In addition to the length of the procurement cycles, regulatory policy in most states results in better economics for ownership of IT in comparison to service- or cloud-based solutions, due to allowed profits traditionally being based on assets. While the National Association of Regulatory Utility Commissioners (NARUC) recognizes this gap and has approved a resolution encouraging level treatment of cloud computing (SaaS) vs on-site hardware, this does not enforce level treatment or solve the fragmented nature of regulatory policy, which continues to be a state-specific matter. These challenges have proven to be major hurdles for even the relatively successful SaaS companies such as Opower, which was acquired for roughly half of its IPO price, and C3 IoT, which was founded as C3 Energy before pivoting to include other markets. Uptake’s revenue model is slightly more complex than other SaaS models, as it involves utility profitability. Uptake has demonstrated that it knows where not to step, and its first step in energy is certainly a promising start.
Uptake’s growth path within the power sector will be very interesting to watch. The IPP-owned portion of the installed wind base offers significant opportunity, though the list of “iconic brands” is, by definition, short. There are also plenty of growth paths outside of improving availability of wind farms. Garg mentions that Uptake may move into solar, which offers a market similar to wind in terms of IPP asset ownership (notably, BHE Renewables also owns 1.4 GW of solar). Uptake could also grow by developing solutions for use cases other than power plant availability, but requiring that use cases have rigorously quantifiable value measurement for non-utility customers reduces the number of viable options. For example, grid operational optimization is an application of predictive analytics that has been targeted by start-ups such as Enbala, Opus One Solutions, and Grid4C; and recently, Google’s DeepMind AI unit entered this space. However, grid operational optimization would not seem to be a fit with Uptake’s philosophy, as value demonstration can be murkier, and utilities are the only customers for this application.
With a collection of brand-name partners and investors, Uptake has generated impressive momentum. While its broad strategy across sectors addresses an attractive opportunity, and the specific strategy within power is designed to avoid well-known pitfalls, growth within power remains a challenging issue. How Garg and Uptake navigate the minefield will be a story to watch as part of the ongoing evolution of IoT.