Putting Agile Personalization Into Practice
by Dave Ingram
This is part three of a five-part blog series on personalization.
Author Benjamin Erwin once said, “Building a robot that works involves building a robot that doesn’t work and then figuring out what is wrong with it.” The software world has used terms like agile and lean for this approach for some time. Fast forward to today and agile marketing has also become a hot phrase. The core idea is simply to break large projects into smaller, more manageable tasks, completing some of those tasks, then pausing briefly to reflect on your progress. Learn all you can, then repeat. If you can put this concept into practice in your own organization, the results will astound you.
Personalization is not exact; a lot of it involves trying different things and seeing how well they work. In order to achieve success with personalization, you must iterate quickly. The more you iterate, the faster you learn. The number of elements that you can base personalization on are nearly infinite, so you need an approach that allows you to tackle bite-sized chunks of work while still seeing meaningful results. We call this approach, which allows your team to start small, learn quickly, and achieve meaningful results as early as possible in the process, Agile Personalization.
There are three main components to this approach. The first is individual tasks or short, meaningful chunks of work, which an agile software developer may refer to as a story. The second is a backlog, or prioritized list of those tasks. Finally, there is a fixed period of time in which you plan to iterate, known as a sprint in agile, but is basically just an iteration, or cycle. These three simple pieces form agile in a nutshell, and when applied to personalization, can help solve many of the challenges faced by marketers today.
A task is a small, self-contained piece of work that delivers value. These are the central units that you’ll be planning with. Generally, it should be something that one person can accomplish in a few hours to a couple of days. A task could be something like “put a new banner on the home page for returning visitors” or “analyze the results of last week’s A/B tests.” These are manageable chunks of work that deliver value. If the first task you do is also the most likely to move you towards your end goals, then you’re going to see returns on that work quickly.
So where do tasks come from? One of the best ways to come up with tasks, especially when you’re starting out, is just to sit down with your team and other stakeholders and brainstorm things you might like to do. Feel free to capture a bunch of ideas that you might not ever do, as you can prioritize them and determine the best ones to spend your time on later.
Remember how you can have an infinite number of tasks? Well, you won’t write an infinite number down, but you need a place to keep all of your great ideas as you have them. There is plenty of technology available for organizing these lists, but for now simply imagine a wall with sticky notes on it. Write down a dozen or so tasks, then prioritize them so that the one you want to do first is at the top, the one you want to do next is right below it, and so on. Prioritizing this list is the responsibility of the Personalization Owner and can be as much art as it is science because you’ll need to balance the priorities of different parts of the business, as well as balance things like expected payoff and level of effort. But don’t let this bog you down. While there are many frameworks and systems for doing this prioritization, just go with your gut to start out with.
If you want a tool to help you do this organization, there are many available. A simple approach is to put all of your ideas in a Google Spreadsheet. Purpose built tools like Aha!, Asana, Trello, or JIRA could all be used for this as well. Remember, don’t worry about the jargon, just get a tool that’s easy to understand and seems to meet your idea of what is easy to use. If your organization already has a project management tool in place that you’re comfortable using, you can start there.
A few weeks ago, I spoke with a woman who works at a large financial institution. They are in the process of implementing a personalization strategy and have spent the past nine months reworking the taxonomy terms attached to all of their content, with no return on investment to show for it yet. Now, don’t get me wrong, having a good content strategy and properly tagging all of your content can have a large payoff when it comes to understanding user behavior and personas and implementing personalization. But there are unquestionably smaller projects that could have been tackled with just a handful of terms, or even personalization that doesn’t rely on past behavior at all, such as some simple messaging geared towards customers on mobile devices or from certain geographies. These are relatively simple to set up and could have proven out the value of personalization early on while other larger tasks are worked on in the background.
Two weeks is a good length of time for an iteration, or what Agile developers would call a sprint. It could be anywhere from one to three weeks, but should not be longer than that. The key to a sprint is to have something tangible and of value to show at the end of each one. This allows you to build momentum, show success to the organization, and feel good about the work you’re doing. It is also a huge help for those of us who tend to procrastinate on larger projects. Most importantly though, it allows us to learn quickly from successes and failures and correct course when the work we’ve done is lacking in any way. For example, if you spend two weeks building some tests for a particular persona or segment and then discover that the messaging that you’re using is not increasing any of your key metrics, you can try different messaging two weeks later and you’ve learned a lot and moved towards value. If you spend three months redesigning and implementing your website, ad campaigns, and email campaigns using that failed messaging, then you’ve lost a lot more time, money, and resources. You also have very little to show for those efforts, except for the learning that you’ve produced, which could have been done in two weeks.
There are numerous benefits to having short, planned iterations, one of which is an extremely valuable concept known as the “retrospective.” That’s just a fancy term to say that at the end of each sprint, you sit down with your team and talk about what worked, what didn’t work, and how you can improve the work that you’re doing. This rapid and regular introspection can help to uncover organization bottlenecks, gaps in skills, process improvement opportunities, and more. It also helps you determine how long tasks take, allowing you to plan more effectively in the future.
Putting Agile Personalization Into Practice
As the saying goes; it was time to eat our own dog food. While theoretically discussing the process of agile personalization is informative, it wouldn’t mean very much if we weren’t practicing it ourselves. While the idea for this blog series and corresponding ebook had already been decided, it still took a couple of months to get started. There was a lot to cover and it was a little daunting. But by putting the subject of the series into practice, rather than trying to crank out the Agile Personalization guide as whole, we broke the work down into smaller pieces of work. Then, each piece was drafted, edited, and published individually, building up to the greater asset down the road. We looked at what we could accomplish in a two-week period and planned out enough of them to get the work done. The idea was to take a big project and break it down to deliver useful and meaningful pieces on a regular basis.
In short, the basic ideas here are simply to break big projects into small tasks, prioritize those tasks, and do small chunks of work on a regular basis—learning everything that you can along the way. As you iterate, you’ll no doubt put a bit more process in place and find other ways to help your team get better results, faster.