Machine-Powered ‘Taste’ - The Future of Algorithms is Curation

It’s fascinating that we still don’t really have a solution to the ‘discovery problem’ – that is, how people find new things online, things they weren’t explicitly searching for. This is despite 15+ years of Amazon’s “people who bought this also bought…”.

In large part, it comes down to the balance between human curation and machine algorithms, a topic that has been on many people’s minds lately:  

  • Ben Evans wrote about search, discovery and marketing to highlight the fundamental truth of search: that it can only ever return you things you were looking for

  • Ben Thompson wrote about curation and algorithms to highlight the strategic evolution of Apple Music as well as news products from Apple, BuzzFeed, NYT, and Twitter

These articles come at a time when we’re seeing Google push into the arena of answering questions you haven’t even asked, with Google Now. In parallel, the rise of curated streams from Apple, Twitter, and media companies is attacking the problem from the other end.

Back when we all just read newspapers...  (Image credit: Willi Heidelbach)

Our take on algorithms at Distilled

One of the most general ways of describing our core competency at Distilled is marketing in an age of algorithms. At other times, we have described it as sitting at the intersection of technology and creativity. We seek to understand both the technological requirements of (for example) getting your website indexed and ranked in Google as well as the insights, stories, and creativity that truly interest humans. At least until the robots start buying things, we need both.

This combination has served us well in the evolution from an era when the most common failure we saw was all-Flash websites, through a time when every website needed more links, to a machine-learning-driven era when we spend time telling prospective clients that it’s not about those links and that they need to do real marketing online.

By understanding algorithms – such as Google’s original PageRank, Facebook’s newsfeed algorithm, spam filters, filter bubbles, and recent machine learning advances such as Panda and Penguin – we gain an advantage over less technical marketers who make incredible campaigns that struggle to get seen, to get found, and to make a lasting impact on their clients’ businesses.

Conversely, by directly acknowledging the power of insights, stories, and creativity, we seek to outperform those who would have you believe that ranking in Google, or getting 100,000 likes on Facebook is a mere matter of rote application of basic technical concepts.

Curation shows us what the future of algorithms looks like

To return to the core points raised by the two Bens (Evans and Thompson), I believe that:

  • The algorithms we have are largely invisible to most people (ask the average person if they think that their Facebook feed contains everything their friends share)

  • Curation is a critical part of making digital services (which covers pretty much everything, since mobile is eating the world) more human, more enticing, and more engaging

  • BUT human curation is not the end-game. All of the current crop of human-curated elements will be taken over by the machines eventually (perhaps in the next 5-10 years, at a guess) – but in ways that users notice barely if at all

Stratechery’s Ben Thompson says:

It’s possible that algorithms will one day be superior to humans at both of these functions, but I’m skeptical: the critical recognition of context and creativity are the two arenas where computers consistently underperform humans.

I believe that this is an example of what I’ve been thinking of as the strong AI fallacy. The concept of strong artificial intelligence (AI) is the idea of a machine intelligence that is “human-like”. Despite all of the advances in machine learning in the last couple of years, this appears to still be some way off (if it’s even ever going to be the answer). However, we are discovering that computers can be better than us at all kinds of surprising tasks without being at all human-like. Google’s self-driving cars suck at chess, but they will be better than you at driving in the next ten years.

In the same way that “search” used to be something a human did by looking through a (human-curated) directory, but was ultimately beaten by computer algorithms, I think we will see the same evolution from human-powered to machine-powered in these areas of “taste”.

The first beach-head in marketing via these curated platforms will be to appear in the (algorithmically-generated) dashboards the humans are using to discover the gems in the firehose. How do you appear in the places BuzzFeed curators do their research? You can bet that they aren’t reading everything anybody writes anywhere.

They will have a dashboard designed to surface interesting discoveries – the precursor of full algorithmic curation – so you need to burst through their bubbles.

Interested in technology and creativity? We’re hiring

Our SearchLove conference in London

All of this context is why we typically underweight prior experience in our hiring and instead exclusively seek smart, curious people with hustle and drive; we know that the marketing of the future will look different to the marketing of the past.

We want to find the people who will help us discover the things that work, implement them for our clients, and share them in our writing and speaking. If that sounds like your dream job, we’re hiring in all three of our offices (London, NYC, and Seattle). And if it sounds like one of your friend's dream job, then you can get an Apple Watch if you refer them to us.

Get blog posts via email

About the author
Will Critchlow

Will Critchlow

Will founded Distilled with Duncan in 2005. Since then, he has consulted with some of the world’s largest organisations and most famous websites, spoken at most major industry events and regularly appeared in local and national press. For the...   read more