The On-Demand Stack

This is an unprecedented time in technology platforms and deployment. One of the biggest consumer trends (among others) to emerge from the SF Bay Area is the idea of “The On-Demand Economy.” Lately, the space has gotten so hot (and some may label it frothy, even), that the term “on-demand” has been loosely used to describe an even larger phenomenon — order everything from your phone, even if it’s not truly “on-demand.” As a result, while the phrase “on-demand” sticks, I like to point out there are two types of species in this genus: (1) services which are truly on-demand, like Uber, Lyft, and Postmates, where users demand a good/service, which then triggers the system to fulfill that demand; and (2) services which allow the consumer to transact by phone, but the service is delivered at a scheduled time, like Instacart — here, the consumer experience feels close enough to on-demand that it all gets wrapped up into one moniker.

How We Got Here (Briefly)
I’m going to quickly gloss over some big topics, just for context: After the 2008 crash, there was a little shift in the economy. People will argue that in the resulting slow recovery and then tech boom, everyone is doing a little better, but many people haven’t gotten FT jobs back (often stopping to look for new work) and have settled into the life of a 1099-er. This has been furthered by the unintended consequences of reforms in healthcare law, which have motivated many employers to cut back their workforces’ hours to be under 30 hours/week, the threshold where employer-mandated care kicks in. And, finally, you have the still unbelievable ubiquity of little computers proliferating across our homes and pockets, all connected to data or WiFi. Those phones, as I’ve written about many times before, form to collect the greatest consumer technology market we have ever seen — yet, mobile distribution remains choked for many, forcing some daring entrepreneurs to monetize by using the phone to aggregate demand and fulfilling that demand through offline operations.

The On-Demand Stack
When we think of “On-Demand,” we think of apps like Uber, and there are many — and more in beta or seeded that may seep into the mainstream. But, if we peel back the layers of these apps, what we find is a complicated stack of business operations, supply, and technologies. I won’t be able to capture them all here (though, please do reach out and I will add to the list)

  • Labor: The apps we use need people (for now) to carry out the labor. Here, startups (such as Task Rabbit, Homejoy, Shiftgig, Wonolo, etc.) work to provide labor in specific verticals where the end customer is a network of businesses.
  • Mobile Phones and Networks (iOS, Android): Sort of obvious, but bears repeating that the proliferation of phones makes this all possible, even despite mobile distribution being hamstrung for many.
  • Infrastructure & Services: What makes some of these apps run, perform, and adapt to users’ needs? Garry Tan had a great tweet last year, something to the effect of “The best software is invisible.” Well, if we looked underneath the Uber app, we’d find a range of technologies, some lumped into SDKs, some built in-house, and some external services that, combined, help with optimizing routes (Addy, Narvar), passenger & driver matching, pricing, fare-splitting, SMS communications (Sonar), promotion/referral redemption & management, app indexing, deep links (Button, URX) and one day loyalty. As labor comes through their funnel, potential drivers need to go through background checks (Checkr, Trooly, Onfido). Of course, when you get into an Uber, you can now integrate your Spotify experience into your ride. I’d also expect more of these complimentary, ancillary services to help delight users and drive deeper loyalty and engagement.
  • Deployment Application Layer: These are apps like Uber and everything else we put on our phones. Too many to count here, of course…

As a final note, I’m both biased positively on this trend (though there are risks) and heavily invested up and down this stack, mostly at the application layer. I do not know how far the trend will pervade outside the early-adopter tech centers and beyond transportation and food. I’d be lying if I said I did. What is clear is that for now, real companies can be built as services for the larger app companies, and I do believe even in transportation and food (the best two daily active use cases) we will see more version and plenty of healthy competition, collaboration, and eventually consolidation. Be careful what you demand.

Haystack is written by Semil Shah, and is published under a Creative Commons BY-NC-SA license. Copyright © 2018 Semil Shah.

“I write this not for the many, but for you; each of us is enough of an audience for the other.”— Epicurus