“I think you always should choose your battlefield,” he said when asked about competing with Uber. “I think to basically compete with a company like Uber, exactly in the same countries, exactly in the same model, would be probably very challenging.”
“We basically build our own model. Choose a model that has network efforts more within a country, instead of globally. Secondly, not too cash intensive. Thirdly, a lot of other factors come into play and I think we try to avoid battlefields that are too intensive,” Samwer added.
This strategy makes sense if your goal is to build a strong local presence that you can sell to your globally acting competitor who is trying to get a local foothold.
I don’t believe this will work well in direct competition with bigger companies like Amazon, Zalando, Alibaba or Uber. Because here’s the thing: A global online retailer can exert network effects in more dimensions than locally confined companies can. (besides plain scale effects which also come to fruition here)
Let’s play it through.
Amazon for example is building out Amazon Prime as a means to create loyal regular customers. Prime works as a big giant bundle, which is more than just faster delivery. One main part and driver of new Prime subscriptions is Amazon’s Prime Instant Video. Here a global strategy gives Amazon quite a lot of advantages. Video like any content business is defined by high upfront fixed costs (production cost or licenses) and very low marginal costs. Hence this becomes economically more attractive the more it is spread out to as many customers as possible.
So: Amazon produces original TV content once (high costs), and then makes this content available in all markets where Prime and Prime Instant Video are available. The upfront costs are spread out globally. The more markets Amazon can do this in the better for Amazon.1 This way, Prime becomes attractive in a way that is hard to copy. Now Amazon is in a position with Prime to offer services and use cross-subsidies across its services to attack companies in a manner that is hard to counter.
Now, this obviously doesn’t mean Amazon is automatically going to win in every market it is moving into. But it means that a Rocket startup, for example, that is going only at local network effects within a country can easily be at an disadvantage against Amazon or other global online retailers.
Or, in different words: Amazon can have a go at these local network effects as well, but the Rocket startup can not create its own huge Prime bundle.
We talked about this with Delivery Hero and restaurant delivery within Amazon Prime:
I have long been sceptical of Rocket Internets approach of building up startups worldwide without even an optionally available platform/service one level above that could offer types of consumer facing network effects. (They have scale effects operations- and technologywise.) It leaves them fulnerable to something like Amazon Prime. Restaurant deliveries can become an example of this.
Rocket Internet may have the ‘global leadership in online takeway’ (see their PDF presentation) today. And its “Global Takeaway Group” with DeliveryHero thrown into the mix is certainly impressive for now.
But are there any moats being built here?
Making restaurant delivery a part of Prime gives Amazon a way to calculate very differently compared to companies which have to make a direct profit from the deliveries. (Remember also that every single part of the bundle serves as advertisement for the whole thing.) This gives Amazon an advantage especially in a two-sided market as is here the case. Namely Amazon can simply offer restaurants a better deal and in the process at the least(!) destroy any attempts by delivery startups at creating exclusive partnerships with restaurants.
Similar arguments can be made about Amazon’s increasingly strong marketplace.
Zalando as a platform can also be a threat to local online fashion retailers. Once brands establish relationships within the platform and integrate its services an international roll-out becomes easier for Zalando, the platform provider. The reasons why a platform approach makes it easier to create a richer offering in markets you already operate in are the same reasons a roll-out into new markets becomes in some respects easier.
That, actually, is why tech startups and investors love platforms: they are highly scalable.
This is also why Oliver Samwer’s note on Uber is, to be frank, nonsensical:
“I think to basically compete with a company like Uber, exactly in the same countries, exactly in the same model, would be probably very challenging.”
Of course Uber is going to be in every country in the very near future. (With similar reasons to what we talked about with Amazon Instant Video.) And Uber, by being the biggest player in this market, has an advantage when it comes to global partnerships with, say, airlines and airports and integration into third-party apps. Being first in some countries won’t count for much in the medium term, if that is all there is.
So, Rocket’s Global Fashion Group might be in for a surprise once Zalando starts entering markets the group’s startups operate in. Especially in fashion a strong global consumer-facing brand (coupled with localized mobile apps and apps based services) can be a strong asset.
Zalando first has to figure out its fashion platform. But once there are enough insights into market dynamics generated at Zalando to create a loose years long plan, all hell may break loose in the global fashion world.
There is a reason why we talk here so much about the strategies of Amazon and Zalando. Both are operating at a scale that is different to most of their peers and both are following unique strategies that give them additional leverage by creating asymmetric competition.
The amount of market influence they will exert a few years down the road can hardly be overstated. This is understandably inconceivable to a lot of people because businesses at the scale we are going to see were impossible pre Internet.
It might be doable in single cases, but I doubt this can be recreated locally confined in most categories at the scale Rocket Internet needs to be successful.
- Optional costs for localization are marginal compared to the initial production costs. ↩