Prerit, this is a great profile of how IndiGo disrupted the air travel market in India. I’m surprised that it’s operating model helped it fend off competition, however. After its competitors began losing market share, I’m surprised they did not try to alter their operating models to achieve even more on-time flights given all the customers that IndiGo captured. One reason must be that IndiGo has only the one type of plane and can be more efficient in ongoing maintenance due to the age of the fleet. Another might be that IndiGo targets the Indian middle class (as TK’s comment pointed out above), and this customer segment does not want the same amenities available on the Jet Airways flights, for example, so IndiGo can accomplish shorter turnaround times for its planes.
Akash, this is a great overview of how Lending Club creates value in unsecured consumer debt. I think an important part of the operating model that ties with this is that Lending Club uses its IT tools to identify types of accounts that might appeal to different types of investors, based upon their risk appetites and the lender’s profiles (reducing the search and information costs for the lenders). I’m interested in how Lending Club’s model would weather an economic downturn because I’m not sure what the differences are in the risk profiles of these borrowers versus borrowers that would go to a bank for the larger loan and longer lending process that you discussed.
On the first question, utilities generally charge a fixed percentage over the cost they incur to supply electricity to a grid, helping those utilities fund consistent dividends for their shareholders (usually over 75% institutional mutual fund type of investors). Therefore, the utilities use demand charges to compensate them for the variability of the customers’ consumption. That said, adding predictive power for utilities is definitely valuable, which is what I think Stem will eventually monetize as it sells data to utilities. It’s a good question whether utilities will pay for the reduced variability to an even larger extent than forfeiting the demand charges.
On the second question, Stem’s software measures the energy usage each month, making the usage transparent and recorded in real-time and viewable online. At the customer’s peak power, Stem would be drawing power from the batteries but it also has current data on the demand charges the customer would have paid. I agree that this system can’t simply share a number and the customer needs to trust its analytics and the process behind it.
Thanks, Prerit! I responded to your questions one-by-one below.
1) In the U.S. grid system, utilities need to purchase all the power that they need for their regulated territory. In the event that they’re coming close to the capacity of the power plants that they have supplying them, those utilities need to purchase energy from the spot market, which consists of higher cost plants that are turned on when everyone has their air conditioners on during the summer (etc.). The penalty for “overdrawing” as it exists in India doesn’t exist in the U.S. since the utilities have to buy everything that their customers are using and there is at least some plant, albeit more expensive, that can supply that need. That said, during moments of peaking power use (everyone has their A/C on, etc.), utilities do frequently have issues due to failure to maintain grid infrastructure (power lines that have more current running in them get hotter, and therefore start dropping and can be short circuited if they hit trees, putting additional stress on other portions of the grid).
2) Yes, large commercial customers frequently sign power purchase agreements directly with power generation companies. This is specifically legally sanctioned in 25 states, Washington D.C. and Puerto Rico. In those cases, commercial customers manage their electricity needs — and, if they need to, will purchase in the spot market to supply electricity during times of peak power. Managing that is a huge part of what Google’s energy procurement team does (given all of their data centers). I agree the analytics could be valuable as a separate business line, independent of the batteries, so that commercial customers can gauge their energy needs.
Thanks, Aravind! On your questions, (1) given in part the agency problem that you brought up, no, the customer doesn’t pay an actual percentage of their energy savings. The agency problem is mitigated by the Stem software control system — the customer doesn’t need to take any action to realize savings. It’s more a “plug and play” technology. The nature of these commercial & industrial customers is that their energy use patterns involve many “spikes” of peak power consumption during the day, so the storage system simply evens these out. The financing contract with the customer would include provisions that would be relevant if the customer were to go out of business (i.e. security in the underlying storage assets).
(2) They are bundled together, but Stem is agnostic in the storage system actually used. In other words, when Tesla’s gigafactory is in full production, as long as their batteries are more efficient, Stem would have no problem substituting those into the Stem package.
Thanks for your post, Lee. As you pointed out, Lithium Ion degrades with time – although these types of batteries would not degrade as quickly as a standard iPhone battery, for example. The use case here for the Stem batteries might not even require the batteries have full capacity. Monitoring the batteries for their capacity is something that the Stem software constantly does, and the periods that we’re talking about for replacement are over 10 years.
The Tesla ecosystem is truly unique among auto OEMs and I like how you characterized its features. On the protection of Tesla’s innovative patents, Tesla has made its patents open-source (https://www.teslamotors.com/blog/all-our-patent-are-belong-you). This move shows Tesla’s mission-driven focus on converting the auto transportation industry to EVs, but it also shows how EVs rely upon network effects (for charging stations, as you pointed out) and also in shaping driving habits.
You laid out clearly how residential rooftop solar can be cost-competitive with retail utility rates, so I think you showed how that business model operates. I think an important part of SolarCity’s operating model is the financing component, as SolarCity’s principal value-add prior to the vertical integration moves you discussed is providing financing for customers. That financing depends upon the availability of the capital to support lower cost financing. Additionally, SolarCity has significant cost (and rising in recent years) of customer acquisition. As the graph above illustrates, the cost of sales has been steadily increasing since 2013 ($0.45/Watt to $0.64 in 2015). This portion of the operating model seems out of sync with the business model of continuing to drive lower solar costs for customers.
This model reminds me of Dr. Shetty’s Narayana Hrudayalaya model in many ways. I’m assuming that the way to get the cost of the surgeries down from $3000 in the US to $50 for Aravind is made up of many factors and I think you hit on the key ones.
Being able to perform 6-8 surgeries per hour is one key innovation in the operating model, it seems, and that would create analogs in cost of $3000 in the US vs. $300 for Aravind. That $300 plus the cost of the lens (bringing the average comparable cost up to $400) puts the figure on par with the average cost of a surgery in India, so I imagine the chief differences between the US average cost and Indian average cost are related to labor and capital costs. The assembly line process avoids process inefficiency to achieve the faster (6-8 surgeries/hour) model; outreach programs appear to accomplish the challenge to keep Aravind’s facilities filled so they have high utilization rates.