Episode 2: The Personalization Team

Welcome to Episode 2 of The Personalization Podcast–The show about understanding, planning, building and analyzing Personalized Experiences. We talked last time about what personalization is, so today we’re going to talk about the very first step you need to take to get started. This doesn’t involve buying software, building complex infrastructure or anything of that sort. Step 1 is all about People. In fact, thinking about what it takes to be successful in personalization, I think there are three things that matter: People, Process and Product. The software itself is last in line because if you don’t have the right people and an appropriate process in place, you’ll be set up for failure no matter how much you spend.

Before we dive into the news, in the last episode I promised a Q&A section, but never gave a method to reach out. So if you want to ask questions, send feedback or just generally get in touch, you can go to personalizationpodcast.com, or tweet to @p13npodcast. P13n is an abbreviation for Personalization. In other words, there’s a p, 13 more letters, then an n. It turns out personalizationpodcast is too long for a twitter handle.

In case Twitter isn’t your thing. You can also find the show on facebook by simply searching for Personalization Podcast and post questions or feedback there.

In the News

The Contextual Marketing Imperative

This paper came across my Twitter feed the other day, and while it was published in October last year, it still seems worth sharing. In fact, it was an infographic which had initially caught my eye, which has a number of interesting statistics about personalization based on the Forrester Survey. Some key stats follow:

  • 91% of marketers are prioritizing improving customer experience through personalization over the next year
  • 66% of consumers reported that personalized offers and content has had an impact on their decision to purchase a product or service
  • 44% of customers say they receive too many offers and promotions
  • Marketers use an average of 15 siloed data sources, resulting in inconsistent customer experiences.

Algorithms need Managers, Too

Algorithms get things wrong and need management.

Defining your goals correctly is one of the most important things that you can do, as this story illustrates:

“Everyone has objectives and directives, but we also know that the end doesn’t always justify the means. We understand that there are soft (often unspoken) goals and trade-offs. We may turn down a little profit today for a gain in reputation tomorrow. We may strive for equality—even if it causes organizational pain in the short term. Algorithms, on the other hand, will pursue a specified objective single-mindedly. The best way to mitigate this is to be crystal clear about everything you want to achieve.

If you care about a soft goal, you need to state it, define it, and quantify how much it matters. To the extent that soft goals are difficult to measure, keep them top of mind when acting on the results from an algorithm.

At Google (which has funded some of our research on other topics), a soft-goal problem emerged with an algorithm that determines which ads to display. Harvard professor Latanya Sweeney unearthed it in a study. She found that when you typed names that were typically African American, like “Latanya Farrell,” into Google, you were shown ads offering to investigate possible arrest records, but not when you searched on names like “Kristen Haring.” Google’s hard goal of maximizing clicks on ads had led to a situation in which its algorithms, refined through feedback over time, were in effect defaming people with certain kinds of names. It happened because people who searched for particular names were more likely to click on arrest records, which led these records to appear even more often, creating a self-reinforcing loop. This probably was not the intended outcome, but without a soft goal in place, there was no mechanism to steer the algorithm away from it.”

The Personalization Team

Skills Required

  • The Creatives
  • The Analysts
  • The Technical
  • The Personalization Owner

This should be a Cross Functional team as much as possible.

It’s fine if some of these people are from agencies, but the key point is that you need to be iterative. Without iteration, you won’t see benefits quickly enough.

For this section in full written form, check out the blog post on this topic.

Learning from Agile

Next episode I’ll do a breakdown of what Agile Personalization is really all about, but the preview for today is that being Agile means iterating quickly over a lot of ideas and properly prioritizing those ideas so that you’re doing the best things first. On top of that, you see results quickly. Quick results mean happy stakeholders and more success.

A couple of reading suggestions: - Lean Startup by Eric Ries - Scrum by Jeff Sutherland

Tip of the Week

Free A/B testing! Google has A/B testing baked into Google Analytics. If you’re looking to get a quick start into Personalization and don’t have the budget for advanced software, check this out. It’s called Content Experiments and it allows you to direct people to 2 or more different URLs and see how they respond to each.

comments powered by Disqus