Saturday, January 26, 2013

Alley Oop in Analytics: Multi-Channel Funnels
 From football to basketball, or even soccer, most Americans have a deep understanding of sports. Digital analytics? Not as much. But, by the end of this article, starting with a basic understanding of basketball, you should know enough to understand multi-channel funnels, their place in analytics, and why anyone with a website should care. More after the break.

So, what is a multi-channel funnel and what does this have to do with basketball? To answer that questions, let's start by getting to the heart of basketball. At it's core, basketball is the game of players running down a court, and shooting a ball into the basket.

In digital analytics, there are many analogues that we can start defining: 

The Basketball = Customers
Basketball court = The Internet
Basketball players = Channels
The Basket = The Goal
Scoring a basket= A Conversion
How the players get the ball down the court and into the net while being harassed by the opposing team is what makes it a game. In some cases, a single player (Channel) can take the ball and run it all the way down and score a basket. At other times, the ball is passed among many players before being taking a shot. Understanding how the ball gets down the court is critical to knowing how likely the player is to score.

How the team gets the ball down the court = Multi-Channel Funnel
How likely the player is to score = Goal Conversion Rate

The relationship between the players is key to understanding how well they perform together, and a strong team that plays well together is a good indicator of success in basketball. A great example of this comes from the 1990s Utah Jazz. As any Jazz fan would know, John Stockton and Karl Malone played exceptionally well together. (And not very well apart).

In digital analytics, the same rules apply. In traditional analytics, only the channel that makes the conversion gets credit. This is equivalent to just looking at the points scored per player in basketball - it gives you some of the information, but you are missing out on potentially important details: Who made the assist.

Stockton to Malone = Conversion Path

Let's think through an example: Let's assume that you have three players on the court, a point guard, a shooting guard, and a center. Looking strictly at the number of points scored, you see that your point guard and shooting guard are performing really well with 7 and 9 points respectively, but your center doesn't seem to performing at all - he has no points. So you decide to take him out of the game and replace him with another shooting guard (after all the first shooting guard already scored 9 points). However, after this substitution, everyone's scores plummet. Why? The center was rebounding - getting the ball back out to the players who make the shots.

This example illustrates exactly why multi-channel funnels matter. They are an attempt to understand how customers make it to the goal by not only tracking conversions on each of the channels, but also tracking how the channels impact each other by analyzing the conversion paths. By tracking multi-channel funnels, you are able to more robustly test to see how one path might be affected by changes to it's channels.

So why does this matter? There are thousands of websites on the internet, all vying for the same thing: customers. Websites that are able to understand their channels and conversion paths, by analyzing and optimizing multi-channel funnels, can work to improve their offering of channels and improve their goal conversion rate (Translated: Teams that are able to understand their players and how they play together, by analyzing how the team best gets the ball down the court, can work to improve their team and improve their likelihood to score). With any luck, this will lead to some paths that perform like an Alley Oop - one channel tosses the customer towards the goal, and the next drives them to the goal with gusto. 

For more information on how to get started using Multi-Channel Funnels, there is a great primer on Google Analytics' Multi-Channel Funnels tool here, and a really good webcast on Multi-Channel Funnels hosted here. In addition to the Google Analytics suite of tools, most of the main
  1. "Gerald Green... Are You Kidding Me?!"
  2. "Basketball"
  3. "Multi-Channel Funnels"
  4. "NBA - Top 10: John Stockton and Karl..."
  5. "Multi-channel Funnels Webinar..."

1 comment:

  1. This is an interesting perspective to multi-channel analytics. Having written two posts on this topic, it was interesting to read your way of explaining the same.