The Case Against Everything on Demand
Delivery startups are cropping up everywhere, which is great in the age of Everything-as-a-Service. But why are they all jumping on Uber’s black on-demand bandwagon? Being able to promise hyper-fast deliveries is neat, but at what cost? Most on-demand startups are probably running on losses and battling a constant last-minute scramble day in and day out.
So before you promise on-demand, let’s consider the alternative of scheduling your deliveries.
Every Silicon Valley success story triggers the birth of many derivative startups. In the age of Uber, everyone is fighting to become the next “Uber-for-X”. Re/code called it the instant gratification economy. We are getting lazier, more demanding, and want everything now. Need a cab? ASAP please! Hungry? Food! Dirty laundry? You get the picture.
However, not everything needs to be available on-demand. It makes sense for taxis and fast-food, because they’ve always been on-demand. But do you really need laundry pickup on-demand? A keg delivery right now? Groceries within the hour when you think of it? A massage? Who has hour-long gaps in their pre-planned calendars these days to allow for a massage on-a-whim? I’d much rather schedule any of the above at a convenient time.
The main reason for this is that “on-demand” does not imply “instant.” You still have to wait for the person to show up, which can take up to 2 hours depending on the service you’re using and how busy they happen to be.
This uncertainty itself creates a waste of time. Think of scheduling as booking a table at a fancy restaurant, as opposed to putting your name on the waiting list “on-demand” and having to stick around until they call your name.
On-demand headache, anyone?
Promising 30-minute service is pretty cool, but exponentially harder to pull off compared to scheduled deliveries. We’ve witnessed the madness and had many conversations with on-demand delivery businesses. This first-hand account on Hacker News sums it up.
Scheduling your deliveries means that you know your workload in advance, avoiding a real-time logistical nightmare. The direct results of avoiding the latter pay significant dividends: happy employees, loyal drivers, better customer service.
Thus far, we’ve covered the softer arguments against on-demand. Let’s talk economics and look at it from a business perspective.
Economies of scale
Profit is revenue minus cost. Your revenue is a function of the total number of deliveries. So are your costs per delivery:
In order to stay profitable, you will need to charge enough to cover your costs. But it is a delicate dance between price and demand. If you charge too much, you won’t have enough demand. If you charge too little, you cannot cover your costs.
Traditionally, logistics providers and couriers have been able to charge as little as they do thanks to consolidation leading to economies of scale. Startups are at a disadvantage here. Size does matter in this market, and it’s really hard to compete with the likes of Uber, who is starting to move into the delivery of things as well.
In order to deliver on your promise of on-demand, you have to have enough resources roaming the city to meet market demand. At the same time, you better have enough volume to keep everyone busy. Otherwise, you’re probably bleeding cash.
On-demand requires more drivers, higher wages (to compensate for the headaches), more distance traveled and higher fuel consumption (because you’re constantly chasing after demand), all of which contribute to an increased cost of delivery.
Scheduling deliveries can generate the same amount of revenue with less drivers, lower wages, and less distance traveled. It dramatically increases operational efficiencies, and hence profits.
Operational efficiency and asset utilization
In the end, it’s about maximizing asset utilization. How many deliveries can you do per driver? Depending on how much you charge for delivery, each driver has to do x deliveries a day to break even. Either way, your objective should be to maximize x.
Knowing in advance where you need to go, will by definition put you in a better position to maximize x.
Another advantage of scheduling is that it allows you to spread the load. Many might prefer their groceries to be delivered in the evenings, but you can offer them a morning delivery for a discount. Peaks in demand are always problematic, even if you have fancy prediction algorithms. Why predict and pray, if you can know for sure? Just call in an extra driver to help with your busy day tomorrow.
Scheduling also enables you to leverage route optimization algorithms. They can automate your dispatching process, while truly squeezing the most out of your fleet — maximizing x — and keeping your business profitable.
The bottom line
If you are building a delivery startup, you can either bet on massive, sustained, and predictable growth, hoping to reach a planned level of scale to break-even before you crash, or you can choose to operate efficiently right off the bat.
Many delivery startups seem to be betting on the former, raising giant VC war chests in the hopes of eventually building a sustainable business. Kozmo raised $250 MM, but was bleeding $29 MM per year when they filed for IPO. Webvan raised $325 MM in an IPO and went bankrupt two years later. FreshDirect, on the other hand, focused on profitability instead of rapid growth and is now a massive sustainable business. 
Scheduling allows you to bootstrap a profitable business, letting you start out with a small customer base and grow organically and smoothly. The benefits of scheduling are analogous to those provided by elastic cloud computing: just scale up your costs when you need to. Deploying a large fleet of drivers is like building a bare metal server farm then trying to fill up the capacity; who still does that?
Don’t get me wrong: there is — and always will be — a market for on-demand deliveries, but I’d argue that the market for scheduled deliveries is at least as large. The big difference is that the latter is much easier to execute on, with higher profit margins and fewer headaches. Scheduling allows you to build a sustainable business, in every sense of the word.
This article first appeared on Re/code
 Even in the food delivery market, there are plenty of customers (mostly businesses) that will know that they need meals delivered a day or at least a couple hours in advance, which also allows for scheduling.
 This is of course assuming that you are not outsourcing your deliveries. Running your own fleet is not easy (especially on-demand) and may not always make financial sense. However, if you want to maintain full control over the customer experience, you’ll need to be running your own fleet.
 To be sure, things are slightly different today. Webvan was trying to build $30 MM warehouses across 26 cities. The current generation of on-demand startups are not as vertically integrated, and maintain relatively lower costs. But still, we are seeing the same patterns of grabbing marketshare at all costs, instead of building a scalable and profitable business first.