Why do we need vehicle data hubs?
[Brussels – 5th April 2020]
PTOLEMUS recently updated its global automobile market estimates and forecasts. We actually found that there are now only 160 million cars with embedded cellular connectivity on the planet, or only 14% of all cars in use!
Why is this so low? Didn’t GM launch OnStar almost 25 years ago, in 1996? Didn’t BMW launch ConnectedDrive 12 years. ago, in 2008? For anybody who has been in this business for a long time, this is depressing!
Causes abound but in our view, the most important reason is that car makers haven’t found a way to generate a sufficient return on their investment in telematics.
Why not?
Well because most of the data generated from this connectivity is typically used by a single entity, the manufacturer itself! Yes, all the investments and operating costs of bringing telematics to cars is supposed to be recouped by the automotive OEM alone. Which means only high end cars are equipped as a standard.
This is exactly the reverse model to the App Store formula where smartphone OEMs are successfully spreading their data to millions of third-party developers applications, daily.
What OEMs have done so far is go for one-to-one data exchanges with stakeholders they deemed large enough to be interesting. For example, GM has partnered with large insurance companies such as Progressive Insurance, State Farm and Liberty Mutual in the US. In Europe, Renault has partnered with Groupama and BMW with Allianz.
However, the number of drivers using services with these one-to-one data connections remains miserably low, which is making any additional agreement a loss-making exercise!
Faced with these failures, OEMs have started to adopt alternative strategies, either building their own vehicle data hub (VDH) or partnering with third party hubs.
For example, BMW launched BMW CarData, its own data hub in 2017. And GM and Daimler partnered with companies such as Caruso, LexisNexis, Otonomo, Verisk and wejo.
What is new?
The car data hubs, sometimes also called data market places, bring scale. As shown below, the number of links required for 4 OEMs to connect with 4 vertical market players (eg. leasing companies or breakdown assistance companies) goes down from 4*4 = 16 to 8.
Let’s now consider the top 10 13 OEMs in the US, which cumulated an 98.7% share of the car market in 2019. Let’s assume that they will need to connect to around 15 stakeholders in 10 vertical markets. In the Vehicle Data Market Global Study, we considered payments (tolls, fuel and parking), insurance, fleet management, car sharing and rental, roadside assistance, repair, traffic information and advertising.
We now have 13 OEMs to connect with 15*10 = 150 third party stakeholders.
To connect these in a point-to-point model, OEMs will need 13*150 = 1,950 links. Creating each link could cost $50,000 on both sides of the equation (for the contractual negotiation and implementation). So the effort would cost $195 million to these stakeholders, only to set up these connections!
Comparatively, in a hub model, with a single data market place, only 28 connections will be required, costing only $2.8 million. Even in a market with 5 distinct hub suppliers, the cost would only be $14 million, only 7% of the cost of a point-to-point model.
Now let’s use the real statistics. There are about 300 auto insurance companies in the US only!
So connecting all key OEMs to all insurers would require 13*300= 3,900 links in a point-to-point system, costing a staggering $390 million. A hub model with 5 market places would require only 1,565 connections. A hub model with 2 market places 626 links.
If we now understand that there will surely be a long tail of developers who will be interested in accessing car data, we can see that the point-to-point model doesn’t scale.
The hub model is the only one that can scale across the planet and across the different vertical markets. This explains why many car makers are now rushing to implement VDHs.
In the 120-page free abstract of our new Vehicle Data Market Global Study, we detail which datasets can be retrieved from connected cars and the level of supply and demand for them.
Frederic Bruneteau
###
For more information, please contact Frederic Bruneteau, Managing Director, at [email protected].