Startup general interest

Service discovery in conversational commerce – might not work for startups

By February 10, 2016 2 Comments

There’s widespread excitement in the startup community about ‘conversational commerce’ – a new shopping paradigm where we buy things virtually through chat interfaces, probably inside the major chat apps – Facebook Messenger, Whatsapp, Snapchat, Slack, Telegram, WeChat, Line etc. Now that traffic to the top four western messaging apps now exceeds traffic to the top four western social networks it is only a matter of time before all the commerce ideas we’ve heard about at Facebook come to messaging apps (not least because two of them are owned by Facebook).

The big question for me in all of this is how service discovery will work. The video above shows how the Uber integration with Facebook Messenger works – users learn that clicking on an address will bring up relevant services, one of which is Uber. Overall I think there are three options:

  • App Stores owned by the messaging company – this is the most obvious, and WeChat in China is already making this work, with over 10 million apps (although they are not apps in the native code sense)
  • Auto suggestion based on parsing what users write
  • ‘Expansion’ buttons which users press when they want relevant services

The worrying thing for me is that in all of these scenarios the messaging platform gets to play kingmaker. Without promotion in the ‘app’ store or being chosen for auto-suggestion or the list behind expansion buttons users won’t find out about services. It seems to me that in this scenario the kingmaker takes the lions share of the upside. Not good for startups.

Some commenters have the view that we will use messaging apps as a sort of command line for our lives, summoning services by writing in code. I get that would be a way around the messaging platform dependency point, but I can’t see it myself. I know that Slack operates a bit like that, but only for some users and I can’t think of a mainstream service that has required people to learn a programming language.

Timing is another question for would be founders in this space. For conversational commerce services to work, not only do the APIs to the messaging services need to be open and the discovery problem solved, payments needs to be licked as well. It would be dangerous to launch a service too long before at least one messaging service has a critical mass of users with payment details.