When making recommendations, the Cassandra Predicament is the worst place you can be in. In Greek mythology, Cassandra was a priestess of the Greek god Apollo. While she accurately predicted fateful events, such as the fall of Troy, she was never believed. Likewise, you may have done your research, analysed the data, talked to the experts, drawn your conclusions and put your crystal ball back on the shelf.
Now it’s time to tell people what you found. It’s time to give them options for action and avoid sharing Cassandra’s fate. I’ve compiled seven principles for making recommendations to help you give the most effective decision advice. Because being right alone doesn’t guarantee that you’re being heard.
What is a Recommendation System?
With the massive growth of online content, users have been overflowed with choices. Recommender systems are designed to understand users’ and products’ preferences, past decisions, and characteristics by analyzing interaction data such as impressions, clicks, likes, and purchases. These systems can recommend various products or services that match consumers’ interests, including books, videos, products, and clothing.
A Statistical Overview
Now, here is a list containing all the statistical data of popular companies that use recommendation systems as a main component of their business:
- YouTube – People upload 500 hours of videos every minute, so it would take 82 years for a user to watch all the videos uploaded in the last hour.
- Spotify – Users can listen to over 100 million song tracks and podcasts.
- Amazon – Users can buy more than 350 million different products.
- Netflix – If a user wants to watch all the series on Netflix, it will take approximately 10.27 years of continuous viewing, assuming no breaks and watching 24 hours a day. It would take about 61.64 years of watching 4 hours per day to watch all the series available on Netflix.
- Facebook – As of 2024, Facebook has approximately 3.05 billion monthly active users, and it’s impossible for anyone to connect with all of them.
- LinkedIn – Approximately 61 million people use LinkedIn to search for jobs every week. LinkedIn hosts over 15 million job listings available for users to apply to at any given time. Additionally, around 8.72 million job applications are submitted daily, translating to about 363,600 job applications per hour or 140 job applications every second.
Ways to Frame Your Recommendations
Before we dive into the principles, it’s worth bearing in mind that your findings may seem obvious to you.
However, they’re most likely a black box to your clients.
After all, that’s why you were tasked to look into an issue in the first place.
That said, your clients may still have preconceived notions of the results of your work and their implications.
If you’re in charge of making recommendations,
there’s also a good chance that you’re not a decision-maker yourself.
But that doesn’t mean you don’t have one or more preferred options based on your findings.
So ultimately, the question is how you can best frame them.
There are five principles you may find useful when making recommendations.
The Gold Standard
Our gold standard is the course of action against which all other options are measured. We can think of it as the intervention we’d pick if the problem at hand was the only one in need of solving. It’s the best option available that could reasonably work under ideal conditions. But let’s face it: When are conditions ever ideal?
The Do-Nothing Option
In my estimation, the Do-Nothing option is too often overlooked. Would I be wrong to assume you sometimes find yourself endlessly zapping through TV channels or browsing Netflix on a Sunday afternoon? Trapped in a perceived obligation to choose, you missed that one button that could solve your problem immediately. It’s usually red and somewhere at the top of your remote. That’s right, it’s the off switch.
Small Decisions
It has become a truism that interventions of any kind tend to have unintended consequences. One of my favourite examples is a billboard ad I once saw in Minsk, Belarus. The government had prohibited advertising alcohol. This caused vodka companies to start producing table water, filling them in remarkably similar bottles. The billboard ad I saw was for their fresh clear water, but everyone knew what it was really promoting. The government’s policy fell largely flat.
Straw Man Proposal
The Straw Man Proposal is derived from the straw man argument. Strawmanning is usually understood as a fallacy that leads someone to attack a position that his or her counterpart doesn’t actually hold. When making recommendations, the straw man can be used strategically to guide decision-makers towards your preferred option.
Pareto Efficiency Analysis
There’s a classic line in The IT Crowd, a British sitcom about a bunch of lazy IT service staff. They’re tasked with sorting out computer problems for technologically inept colleagues. IT guy Roy tends to answer the phone with two questions: “Have you tried turning it off and on again?” and “Are you sure it’s plugged in?” The running gag is that this often solves the IT issue without even diagnosing the problem.
Types of Recommendation System
Three main types of recommendation systems are widely used:
- Content-Based
- Collaborative Filtering
- Context filtering
These recommendation systems rely on feedback data regarding ratings, clicks, likes or dislikes, etc. Before we move to the types, let’s explore the two main types of feedback:
Prioritize Individualized Recommendations
A piece of content clearly picked just for you is always preferred over a generic item. On the web, information that was personalized to the individual was seen as a valuable feature, helping users sift through the vast inventories on ecommerce or entertainment sites to find those few pearls they were interested in.
However, these personalized recommendations were difficult to locate on some sites because they were positioned too low on the page, below generic sections of promoted content. This placement made them less discoverable, rendering their inclusion on the site basically worthless.
Separate Categories of Recommendations
Another benefit to displaying the specific source for a set of recommendations is that it tends to force the separation of personalized content into smaller chunks, rather than having all recommendations lumped together in a single category. Users appreciated being able to browse more specific categories of recommendations, especially for sites with large and varied inventories — whether that inventory was ecommerce products or entertainment content. Just as you wouldn’t throw all the content on a site into one big listing page, don’t throw all recommendations into a single grouping.
What is An AI-Powered Recommendation System?
AI-powered recommendation systems stand at the forefront of modern technology. They employ machine learning algorithms, predicting and suggesting items or services. Users find these suggestions tailored to their preferences.
You encounter these systems daily. They’re behind the recommendations on Netflix, Amazon, and Spotify.
Machine learning algorithms are the heart of these systems. They analyze user behaviour and preferences. Patterns and similarities are identified. Based on this analysis, predictions are made. Suggestions are then offered to the user.
These algorithms learn from past data. They improve over time, making more accurate predictions. User preferences play a crucial role. The system learns from user behaviour.
Likes, dislikes, and past behaviour are considered. The system understands the user’s taste. Recommendations are then personalized. The user finds these suggestions relevant and engaging.
Predictions and suggestions are the system’s output. The machine learning algorithm makes predictions. These predictions are based on user behaviour and preferences. Suggestions are then made.
These suggestions are items or services that the user might like. The goal is to enhance the user’s experience. AI-powered recommendation systems are everywhere. They’re integral to many platforms. Netflix uses them to recommend movies and shows.
Amazon uses them to suggest products. Spotify uses them to suggest songs and playlists. These systems enhance user experience on these platforms. They make the platforms more engaging and personalized.
Bonus Principles for Making Recommendations
Now, it may very well be that you still have plenty of potential options to recommend. You’ll most likely have to strike a balance. Between those that actually solve the problem and those that will be acceptable to decision-makers. Here are two bonus principles on how to identify the least risky option and on the art of persuasion
Irreversible and Reversible Decisions
As our first five principles imply, some options are riskier than others. In order to tell them apart, the distinction between reversible and irreversible decisions can help. It goes back to a shareholder letter written by Amazon founder Jeff Bezos.
Tactical Empathy
“The most dangerous negotiation is the one you don’t know you’re in,” says former FBI hostage negotiator Chris Voss. Put bluntly, if you care about the impact of your proposals, you’re in a negotiation. According to Voss, whose book Never Split the Difference I reviewed earlier, Tactical Empathy is the foundation of persuasion.
Reasons for recommending someone
If someone asks you to write a recommendation on their behalf, you have a choice. You can say no. I recommend you say no if you cannot answer yes to all of these questions:
- Do you have genuine affinity for this person?
- Are you enthusiastic about their abilities, strengths, and potential?
- Do you want them to get into the program, team, or job they are applying for (even if that’s bad news for you because they will leave your team or organization)?
- Are you willing to do the work yourself (and not ask them to write it for you)?
If you answered no to any of those question, then you might be doing a disservice to the candidate if you agree to recommend them.