5 Algorithms every Product Manager need to know about

If you think finding the right technological solution to a problem is at the heart of Product Management, then having a mastery over well-established business techno-logics (loosely called ‘Algorithms’) can differentiate the Great Product Managers from the Good ones.

Here, We are presenting Algorithms that affect user’s life and choices so dramatically, that it is not possible to be a PM without understanding them. Algorithms grow complex with time and scale and thus only a very brief introduction is provided here. For more info, it is advised to go through company’s developer pages and ofcourse Google them!

1. Google Search Algorithms

Famously called the “Page-Rank algorithm”, This is the ‘bread-and-butter’ of Google. Simply explained, when you search on Google, it finds the most relevant top-10 results for your question on the basis of importance of the web-page.

Google uses spiders to crawl over webpages and store them in its database as list of pages called “Index”. The algorithm can estimate a webpage’s importance by looking at other important pages that link to this page. Thus, each web-page gets a score based on every other page that links to that page. Also, the PageRank takes into account the quality of incoming links (such as your website gets mentioned by The New York Times) more than the quantity.

With time, the logic has become more complex with Google now providing for Wiki, words with similar meanings, recommendations, products to buy, Google maps for how to reach (in case you search for a place) etc.

2.Spotify Discovery Algorithms

Every week, Spotify provides a personalized list of top songs that you would like. Their “Discover Weekly Algorithm” for personalized auto-recommendation is considered the best in the industry with most users confirming that songs they get in list are really the ones they like the most.

Source: Quartz

It looks at the songs that you have listened to and at all the playlists that other people have made. It then uses “collaborative filtering” to compare the two datasets to figure out which songs you might like to listen to (for eg. if person A has listen to 8 out of 10 songs that are in person B’s playlist, Spotify may recommend other 2 songs to Person A)

However, it also goes beyond that to build your “User Profile” based on songs you listen to, their frequency, songs you skip- genres, artists, moods. It also takes into account other parameters such as age, gender, location among others to match your profile with others and with their playlists to finally determining the most relevant songs.

3.Amazon’s “Recommended because you purchased”, “Recommended because you added X to wishlist” Algorithms

Though Amazon is world class, renowned technology company, core of its once famous recommendation algorithm is now pretty out-dated.

In case, Mathematics is your thing.

It uses what is called “Item-to-Item Collaborative filtering”. For each item you buy/add to wishlist/look at — Amazon ranks other items on basis of cohorts of other users’ list. The items that get the highest weightage (called ‘closest to neighborhood’ ) are showcased as the relevant recommendation. The crucial part is to define the cohort- do you go through list browsed by all users or a well-defined cohort of users.

Defining this cohort require various parameters including- Location, Purchase history, demographic segmentation etc.

Currently, the most powerful algorithms in “recommendation system” category are by Spotify, Netflix and (believe it or not) Target.

See how Target figured out a teen girl was pregnant before her father got to know:-https://www.forbes.com/sites/kashmirhill/2012/02/16/how-target-figured-out-a-teen-girl-was-pregnant-before-her-father-did/

4.Facebook’s “News Feed” Algorithms

True answer: Nobody knows exactly how this works. Infact,there are have been outrages by users whenever FB tweaks this algorithm- as it defines the most basic and core interactions on the social network. Sadly, the number of relevant posts have been going down since 2017, when FB made considerable changes to its news feed.

There are more than 150,000 factors (personalised or otherwise) that Facebook takes into account. Key ones are:

1. Personal Relationship to user: Post of Person you have interacted with more on FB gets a better ranking
2. Post Score: Depending on no. of comments, type of comments (such as ‘Congrats’ or ‘Great’ means important post), no. of likes and now complex metrics of ‘Reacts’ etc.
3. Type of Post: Emotions raised (love, sad, happy, like etc.), video, photo, article, gifs, total characters in a post etc.
4. Recency: Latest get better ranked than say, 5 hrs or 1 month old
5. Platform of User: If you are using PC, more videos. If using Mobile phone, more photos and texts.

Facebook’s major goal is to improve the relevance of the posts so that user engage more with the product and spend more time on FB. As of Feb 2018, FB also has included many parameters to figure out fake posts- that it ranks severely low.

Bonus: Facebooks’ “Graph Search Algorithm” — Based on NLP (Natural Language Processing), this algorithm was used to power “Search Engine” of FB. (For the old ones here, remember when FB allowed you to search “People who live in “ABC” are “Men/Women” of age “greater than abc”). It was taken down by FB due to privacy issues among others. However, a part of it still used to recommend new friends to the user.

5.Tinder’s “Elo Score” Algorithms

The matching algorithm is at the heart of Tinder. From outside, the app seems simple: just showcase all the pictures of people near the user and if there is a right swipe from both side, its a match!

But no, Tinder’s core metrics is to ensure probability of a match is high (ie. no. of matches/ no. of swipes should be lowered) as well as the matches shouldn’t get concentrated (for eg. you go to a bar, and most guys hitting on the same hottest girl) For this, Tinder uses variety of factors:-

1. New users get a boost- call it ‘Beginner’s Luck’ in digital dating
2. User Attractiveness Rating on scale of 1–10. If a user is rated 6, user will only see 4–8 rated other users. This Rating works similar to “PageRank” algorithm of Google. It depends on the rating of users that swipe you right, as well as with users that you gets matched.
3. Activity status- the more active user is on Tinder, the more matches the user gets, the more messages the user send to the matches- the better is your rating.
4. Paid Account- If you pay for Tinder, your rating is boosted.
5. Pickiness- No. of right-swipes/no. of users showcased — will also determine you ranking or Elo Score.

Bonus: Tinder’s “Smart Photo” Algorithm — Tinder automatically does A/B testing of user photos to increase the probability of it being swiped to right. The feature can track which of the user’s photos get the most swipes, and then puts that photo first. This method also works by tracking a user’s swiping habits in order to optimize which pictures on other people’s profiles are shown first.

In Upcoming Post, we will look at powerful machine-learning algorithms driving Quora, FB, payments services, Uber, Netflix, Amazon among others. But before that, learn the fundamentals of AI,ML and Deep Learning.

Stay Tuned!

You can reach me at: https://www.linkedin.com/in/rds/