The discovery problem.

Paul Kinlan

I just saw:

I thought this was incredibly interesting. There is something to be said for this. It is something that I have been thinking about for a while as we think about web apps, and having users find them.

I have spoke to a number of colleagues about this, some of them have worked in very successful businesses where the discovery problem was a real problem and creating something successful is normally a combination of the following:

  • You need to market your app or site and you need to spend real money.
  • You need to be in the places where users are (i.e, the marketplace)
  • You need to be able to convert a user. This could be from a visitor to a user, or convert to a payment etc.

I think the phrasing is wrong when we talk about discovery, we look at it in terms of keywords, or top 10 lists, or category buckets and SEO. We should be thinking about matching user intent and the marketplaces that can match that user intent. I suspect it is one of the reasons why search advertising works well and converts well: you know the user is searching for a flight, advertise them a flight.

When you think about matching intent it becomes a lot clearer in my mind, people want a good, yet fair system of indexing and ranking based on the intent of the user. Assuming good intentions, a discovery engine (search, or whatever) will automatically rank results based on a set of heuristics to give you a list of items that it thinks matches the users intent in order of quality or applicability.

When I worked on Web Intents, for every action that the user wanted to perform (such as share link) we could either fulfil it directly from their list of installed apps, or you could query the market (Chrome Web Store in this case) and because we knew the user's intent (I need to edit an image) we could, if the market had the goods, fulfil it.

This got me thinking about the marketing part, and in my head I am in two spaces (I am pretty sure there are many more types, and what I am talking about might have a proper name, but this is what I am thinking):

  • Intent based marketing,
  • Blunt marketing

Intent based marketing comes when you are in marketplace and the index matches the intent of the user, your marketing is to effectively try and be more prominent than the natural ranking of what ever the marketplace uses. In the case of Web Intents, I could have quite easily envisaged a market where Adobe would pay to get their new App into the top of the application picker, because they directly knew the user wanted to “edit an image”.

Blunt marketing is for everything outside of that marketplace. I see on mobile a lot today where the marketplace is ill-matched to the users' intent. And whilst mobile advertising seems to be generating a lot of revenue at the moment, it mostly involves advertising apps from within apps, or on TV etc.

I think app store search is broken and doesn't match the results of the index well with the user's intent, which is why we see such a huge market for Blunt marketing. Good apps don't get found by users as well as they should be, so successful apps will normally have a heavy marketing budget behind them.

I have no data (yay!), but I suspect Intent based discovery is a lot more effective and will provide more value for people who want “discovery” AND people who have budget to market. Is Blunt marketing bad? No, but in many cases it shows the failure of the marketplace (and the ability for natural discovery of quality).

How does this apply to web apps? The discovery problem is not going to go away any time soon until we can match the user's intent in a search, or action, and fulfil that request with apps on the web. We don't really have a marketplace, or the goods to sell in that marketplace… yet.

Paul Kinlan

Trying to make the web and developers better.

RSS Github Medium