Coffee, Coffee, Coffee!

In 1984, Robin Williams starred in Moscow on the Hudson as a Russian saxophonist who decides to defect from the USSR during a shopping trip to a New York department store. One scene stands out, particularly to those of us who focus on search experience.

This scene, shown below, depicts the protagonist going to the supermarket to buy coffee. Expecting to find a “coffee line” where he will have to queue up to get a single kind of coffee, he is literally overwhelmed by the abundance of coffee options available to him in the “coffee aisle”. 

Some people interpret this scene as either a celebration or a mockery of American free-market capitalism. But, as someone who works on search, I see a parable about information overload.

Indeed, a search for “coffee” on most grocery ecommerce sites leads to an experience that has me mumbling, “coffee, coffee, coffee!” At least the coffee aisle in the grocery store is organized by brand and style — the search results are rarely so easy to navigate. Instead, they are as disorganized as they are overwhelming.

It’s not that these sites don’t have tools like facets to help searchers navigate the myriad available options. The facets are there, but they are usually tucked away in a side panel — or, worse, hidden behind a button for filtering options. In other words, the faceted organization is not treated as a first-class part of the search experience.

To be fair, not all search tasks are like this. For example, if you already know what brand of coffee you are looking for, a search for that brand probably reduces the results to a manageable number. Moreover, a reasonable ranking model should usually manage to meet you needs on the first page.

But broad and ambiguous search queries require support for the searcher’s journey. And it’s not hard to find these queries — they are the queries where searchers experience the most friction.

So please, if you are providing a search experience — even if you don’t sell coffee — do yourselves and your searchers a favor and stop overwhelming them. Your searchers will be happier — and they will buy more, too.

Leonardo Lezcano

Distinguished Data Scientist - Head of Machine Learning of the Walmart Translation Platform

4 年

Thanks Daniel, as usual, your articles are both informative and entertaining. My position regarding choices stands next to the "Paradox of Choice" by Barry Schwartz. The most durable purchasing satisfaction tends to be the one that requires binary or ternary decisions. Of course, this doesn't mean that options are maxed at 2 (like iPhones successfully do). A sequence of simple decisions can build a tree of options with as many leaves as needed. After so many years of retail though, everyone in the business is probably aware of such principle. However not all retail experiences are based on it. One reason could be the Gruen effect, but I'm sure there are many more...

Edward Fine

AI and Data Science Consultant, Technologist and Instructor

4 年

I find that different people have different views on searching based on 2 properties. One is knowledge of self, i.e. what is your own utility function. The other is the tradeoff between search costs and the desire to minimize ex-post regret. The archetypal decisive executive is good at minimizing search costs, while those who are subject to analysis paralysis are often troubled by fear of ex-post regret. So to drive transactions, help your users discover their preferences and help minimize regret.

hmm... It's true that overwhelming choices are bad under some circumstances: https://faculty.washington.edu/jdb/345/345%20Articles/Iyengar%20%26%20Lepper%20(2000).pdf But I am not so sure that it holds true for search and that facets are a good solution. Only if I know in advance that one of these facets is indeed the filter I would want to apply. But if it is, why not put it into the search box on the first place and for the search engine to be able to correctly interpret my query?

要查看或添加评论,请登录

Daniel Tunkelang的更多文章

  • ChatGPT, Are You Just Telling Me What I Want to Hear?

    ChatGPT, Are You Just Telling Me What I Want to Hear?

    These days, the Turing Test — which Turing originally called the “imitation game” — feels hopelessly outdated. With…

  • Not All Recall is Created Equal

    Not All Recall is Created Equal

    Search application developers constantly navigate tradeoffs, particularly between precision and recall. Precision…

    1 条评论
  • To Bot or Not to Bot: It Depends on the Question

    To Bot or Not to Bot: It Depends on the Question

    I was one of Quora’s earliest users. I earned Top Writer status for several years and even made some money through…

  • Ground Truth: A Useful Fiction

    Ground Truth: A Useful Fiction

    A key concern about AI is that models “hallucinate” — technical jargon for saying that they make up things that look…

    5 条评论
  • Conjunction, Disjunction, What’s Your Function?

    Conjunction, Disjunction, What’s Your Function?

    Like many folks of my generation, I grew up on Schoolhouse Rock, a series of animated educational shorts that aired…

  • Modeling Queries as Bags of Documents

    Modeling Queries as Bags of Documents

    Last week, I had the honor of presenting “Modeling Queries as Bags of Documents” at Search Solutions 2024 with Aritra…

  • Documents, Queries, and Categories

    Documents, Queries, and Categories

    I have published a number of posts and presentations about the bag-of-documents model, which essentially represents…

  • Where Do Categories Come From?

    Where Do Categories Come From?

    In my previous post, I argued that categories are fundamental for search applications. I characterized a robust set of…

    1 条评论
  • Categories are Fundamental for Search

    Categories are Fundamental for Search

    As a search consultant, I have learned to be flexible about structured data. However, I do insist on content being…

    5 条评论
  • Quo Vadis Nunc, Quora?

    Quo Vadis Nunc, Quora?

    I was one of Quora’s earliest users, earned Top Writer status for a few years, and topped the leaderboard as a 9-time…

    2 条评论

社区洞察

其他会员也浏览了