Used automobile marketplaces contributed to a strong September quarter for venture capital funding. The space is incredibly competitive with well-funded players like Mahindra First Choice ($30MM), CarTrade ($240MM), CarDekho ($142MM), Truebil ($10MM) apart from the $200MM Cars24 and Droom have raised between themselves. Why on earth do you have $500MM of funding chasing used cars and automobiles?
Turns out, used vehicles are a $100Bn market, with cars accounting for $20Bn. Used cars turn 3.3MM units, and first hand cars turn 4MM annually. When you have a volume as large as actual car/two-wheeler sales, you’re talking really big numbers. In my piece on Indian re-commerce, I had demonstrated how second hand goods were a huge Indian opportunity. The same is true for vehicles, because the Indian consumer is simultaneously aspirational and value conscious.
Droom, along with its fellow competitors, is selling precisely the dream of owning your dream car/scooter/bike (or even plane?). Droom was started by Sandeep Aggarwal, the erstwhile co-founder of Shopclues, where he had a very public fallout with his then wife and co-founder. While most of us don’t really know what the unicorn Shopclues does, starting another company after such a personal setback is commendable.
“We want to make an entire ecosystem around the car” says Mr. Aggarwal (remember this). But what is the real problem that it is solving in the first place?
In his now seminal paper “The Market for Lemons“, George Akerlof argued that the presence of information asymmetry will result in the second hand car market having only lemons i.e. useless cars. Unpacking this says that the buyer has much lesser information than the seller about the car i.e. information asymmetry. Assuming the seller knows a buyer will pay a fixed price, the seller is inclined to sell faulty cars (“lemons”) and not good ones (“peaches”). Over time, the market gets flooded with lemons.
So what does a third party marketplace solve for here? Trust, or technically – the reduction of information asymmetry.
Third party marketplaces add more information to the car by doing inspections, test drives, maintaining the car, assuring buy back and even coming up with a score. The best way to solve for trust is therefore more data, especially for assets with long life cycles. The longer you track a vehicle, the more certain you are about its value – data is key. (Aside: Second-hand cars are a wonderful use case for blockchain, asblockchain solves for trust digitally).
India is a notorious market for second-hand products. The vehicle market is also not spared and is fairly disorganized, fragmented and offline. There are more than30,000 dealers handling cars alone, and likely many more handling the two-wheeler segment. These dealers are therefore selling 100 units a year (or 8 per month), if you assume 3.3MM units. If you divide this by the huge number of car models in India, these dealers clearly lack access to the benefits of scale in terms of price discovery.
They, though, have quality access to local market and customers wanting to sell cars.
As a matchmaker between a vehicle buyer and a seller (usually a dealer), Droom has become a sales tool for dealers. Droom takes an initial commission of 1.5-2% on the sale of a vehicle, the classic marketplace monetization strategy (like Amazon). According to Mr. Aggarwal, the company has three other revenue streams – dealer subscriptions ($30/month), data tools ($2/report) and advertising (case to case).
If Droom sold $20Bn worth of cars (2% is $400MM), to all dealers ($12MM of annual subscription), with 1MM reports ($2MM) and ads for each car (0.5% of value or $100MM), it’s a decently sized market in India.
But the market is bigger than this. Droom wants to be horizontal (i.e. across vehicles), because you could create a potential $1Bn revenue company. But is Droom really selling a vehicle?
It isn’t, it is selling vehicle data. That opens up myriad ways of monetization (e.g. auto insurance)
Droom, the data driven asset seller, wants to monetize a vehicle over its lifetime (remember, the ecosystem around the car?). The sale is just the first monetization strategy (Droom: 2%). You could originate car insurance, premiums could be 30% of a car’s value over a lifetime (Droom take: 2-3%), have maintenance (Droom take: 1%), ads (Droom take: 0.5%) – easily 5-7% of a car’s value. A car could also be sold twice through Droom, adding another 2-3% and taking Droom’s car lifetime value to potentially 10% of a car’s value. For a $5,000 valued car (could be higher), that’s $500.
For 3.3MM cars in various stages of their life cycle going through a marketplace, it’s $10Bn of actual revenue (not GMV). With two-wheelers included, you’re looking at a$50Bn market. The data play is immense – once you own the vehicle’s data, you can do magic with it. But the market is not without purpoted “fraud”.
“It’s a funding company” writes the Ken in an entertaining investigative piece on Droom. The claim is that Droom pays dealers for selling cars through its platform, and is bloating GMV to raise capital. The other interesting claim is that Droom is delusional about lifetime value because other marketplaces just had bargain hunters with no lifetime value, as they never returned.
Firstly, Droom is actually buying vehicle data. As I mentioned earlier, it can utilize this in a myriad number of ways. For the $700MM GMV, the company is valued at1x GMV (Flipkart was valued at 3-4x), so it’s not “hyper valued” even though it has a lower fee for the first order. Secondly, Droom has acquired a vehicle (and incidentally, a customer and dealer). It doesn’t need to reacquire a customer to make transactions related to the same vehicle to old/new customers. This is unlike other marketplaces, which need to keep selling new products the same customer. Both these make Droom fairly compelling.
The true concern is competition having better data than you on the same car, and this is why the market may actually become an asset land grab (buy data of all cars in the market). You now see why Droom has such detailed data. If you are the sole “owner” of the car data, you can monetize the car until it is useless.
The company’s financials are telling – for gross/net it did $100MM/$1.4MM in 16,$280MM/$4.4MM in 17, $1.8Bn/$45MM in 18 and expects $3.5Bn/$115MMin 19. Dividing net by gross you get 1.4%, 1.6%, 2.5% and 3.2%.
Well, expectedly, Droom’s take rate is increasing because it is monetizing more. Despite the sales-y nature of the founder, the company seems to hit its revenue numbers to 60-70% forecasts (although the founder has been claiming elusive profitability since 2016).
Onto unit economics (which is a vehicle). If the acquisition cost of a vehicle (“Vehicle Acquisition Cost”) is 1% (per what Droom “pays” for data), for an LTV of even 3% of asset value, the company has 2% left. Let’s say this is the “gross margin”. On a per car basis, the only other cost is customer acquisition and Droom can pay upto 2% of $5,000 or $100 (INR 7000) to acquire a customer. Let’s assume it pays 1% to acquire a customer.
As the LTV (as a % of car) value increases, Droom makes even greater profit per car. If done right, Droom could make 4-5% of profit per car, over a lifetime. For 3.3MM units, sold at $5,000, and assuming benefits of previous cars sold on the platform, that’s a $1Bn of profit before other expenses. Given this is a data/tech company, and Aggarwal says “GMV per employee is $1MM“, the company could do really well with technical leverage.
Droom’s data-driven dreams could make a compelling, profitable internet company.