I was at a conference several years ago, sitting in the audience listening to a panel of data scientists from Netflix, Amazon, and Spotify talk about the gains they generally see from their data science initiatives. It was a popular talk, as the topic of lift is one that has always been important to business leaders and practitioners.
During the conference, one of the panelists said something that has stuck with me ever since and I want to share it here with you today to help set expectations when approaching data science projects that target increases in user engagement:
Major product changes have the potential to show a 300% lift in user engagement. Using data science or A.I. to personalize that same product offering to users has the potential to show a 30% lift in user engagement.
That’s a huge 10x (or, if you prefer, 1,000%) difference between what can be achieved by product changes and optimization! And it’s pretty easy to see why this is the case.
Imagine taking a good user experience on a shopping site and making it horrible on purpose by removing, say, all product images. Who would want to shop there, using only product names, descriptions, reviews, and prices? No one. Engagement would fall. From this perspective, no amount of data science could be expected to provide as big a lift in engagement as simply adding product images to the site.
This is an extreme example, but illustrates the value of good user experiences and gives perspective on what we can and should expect from A.I. applied to user engagement.
It’s generally unreasonable to think that using A.I. to optimize user experience by delivering highly personalized content would give huge gains. Incremental gains, yes, but if you want big leaps in engagement (and, hopefully, revenue) combine smart changes to your product offerings with personalized messaging and get the best of both worlds.
Generally speaking, in e-commerce, A.I. should support the product – not be the product.
We’ve seen this over and over with our clients. One client decided to offer their products on a subscription basis, without the shoppers having to log in and re-order each month. That small offering provided huge revenue gains. And data science was right there to support the timing and personalized presentation of that effort, but it wasn’t the driving factor, the option to receive products on an auto-renewing subscription-basis was.
Great products require great user experiences. A.I. can help, but should never be thought of as a silver bullet to recover a poor or inadequate interface. Too many companies think that A.I. can somehow save ailing products and drive huge profits. I believe this is where the “hype” around A.I. comes from.
I hope this perspective helps you set realistic expectations for your next data science project. If you want to talk about how much lift A.I. can bring to your company, hit reply and let’s have a chat.
Have a good week!
– Zank
Of Interest
Google’s Latest Experiment is Keen, an Automated, Machine-Learning Based Version of Pinterest
A new project called Keen launched from Google’s in-house incubator for new ideas to help users track their interests. The app is like a modern rethinking of the Google Alerts service, which allows users to monitor the web for specific content. Except instead of sending emails about new Google Search results, Keen leverages a combination of machine learning techniques and human collaboration to help users curate content around a topic.
https://techcrunch.com/2020/06/18/googles-latest-experiment-is-keen-an-automated-machine-learning-based-version-of-pinterest/
Djay Pro Uses A.I. to Turn Songs Into Acapellas and Instrumentals On-The-Fly
A.I. and machine learning seem like buzzwords at this point, with their mention spattered across press releases and new product launches. The ubiquity may have numbed us, but some cool and weird things are happening! In June, Algoriddim for example launched the “Neural Mix” as part of the new djay Pro A.I. These new tools leverage Apple’s Core ML framework and the A12 Bionic chip (or higher) to provide on-the-fly audio separation. Read more here:
https://www.engadget.com/djay-pro-ai-machine-learning-neural-mix-isolate-audio-elements-on-the-fly-130014409.html
Google Sheets Will Soon be Able to Autocomplete Data for You
Google has announced a couple of updates to Google Sheets that will make building spreadsheets and analyzing data in them a bit easier. The most interesting feature they announced is the upcoming launch of “Smart Fill”. You can think of it as Smart Compose, the feature that automatically tries to finish your sentences in Gmail, but in this case autocompletes data spreadsheets for you. Read more about it here:
https://techcrunch.com/2020/06/30/google-sheets-will-soon-be-able-to-autocomplete-data-for-you/