square peg being forced into a round hole
PHOTO: shutterstock

The term agile is at times thrown around like a brand new approach. That ignores the fact that agile is nearly 20-years-old and has long been used by programmers and other IT people as a methodology to develop software. Agile was first described in a document produced by 17 developers between Feb. 11-13, 2001, who sought to overhaul software development processes that they saw as cumbersome, unresponsive and too focused on documentation requirements. 

According to agilemanisfesto.org, the online home of The Agile Manifesto, the goal was not anti-methodology, but rather "to restore credibility to the word methodology."

Agile Development and Business

Here we are all these years later and agile has also become a business methodology that stresses the collaborative nature of business development and aims to achieve goals through more flexible, agile team-work that stresses the need for horizontal cooperation rather than top-down decision making.

Needless to say, it has also entered the digital workplace and development of digital experience (DX) technology. At this year’s DX Summit in Chicago, it was clear that agile and DX technology are already dovetailing in the enterprise as both the technology and business sides of organizations come together to achieve business goals using this more flexible and collaborative

Phil Kemelor is an independent consultant with over 25 years of applied experience in marketing, analytics and business process improvement. He develops solutions that require the integration of multi-channel data and digital transformation at global brands. He also creates branding strategies for analytics consulting firms and writes content for senior executives seeking to build their personal brand. At the event, he chaired a panel consisting of Jeff Kelley, senior vice president and head of marketing at Calamos, Kate Stewart, product manager at CBC digital products and Stacey Decker, deputy managing editor for digital at Education Week, which looked at how agile is being used to choose DX technology. 

Agile Choices in the Organization

Afterwards, he took some of the issues that emerged from the panel discussion to outline how agile has been evolving in the organization. The starting point, he said, is to really understand what an agile approach really means in the enterprise. When talking about agile we are talking about a way to be collaborative and a way to be user-focused when organizations are selecting technology, he explained. This is somewhat of a departure from a traditional vendor selection methodology, which involves processes that have a very rigid set of technical requirements that do not deal with the actual user scenarios. “By that I mean how people in your organization are actually going to use the technology,” he said. This is the key.

Agile, in fact, differs greatly from the traditional process of technology selection. “One of the way it differs is that it looks to really understand the users — even all of the users — in the workflow and how they interact with the systems,” he said. In practical terms this means understanding what the users' tasks are going to be. When tech and business leaders really understand what users’ tasks are going to be they can really drive and build the system taking those needs into account.

This is in contrast to the traditional way of selecting, building or buying technology. In these scenarios, the purchasing team will be the ones that lead the process and drive it using a set of perspectives that may not necessarily be those of the user.

If all users and all requirements are taken into account, will this not mean too many people in the kitchen?  It shouldn’t really Kemelor said, and won’t be if they put certain safeguards in place.

There needs to be management self-governance in place; there needs to be a comprehensive set of issues discussed before anything happens; there is a lot of upfront work in terms of who is going to be involved in the selection team and clarifying what the section criteria are going to be. 

That way, when you get into the decision making process you don’t get bogged down into he-said she-said scenarios as there is a clear list of considerations that need to be met before a decision is made.

Kemelor does point out, however, that this kind of decision making is time consuming as it demands a lot of time of stakeholders and vendors and may be more suitable for big large scale enterprise buys rather than point solutions where the number of users is small enough.

Developing Lyft’s Digital Experience Platform

Ajay Sampat is engineering manager for growth at San Francisco-based Lyft. Lyft built its Marketing Automation platform, Symphony, to manage a large-scale marketing budget across various channels, to improve cost and volume efficiency, while enabling the company’s marketing team to run more complex, high impact experiments. 

The development of Symphony follows this paradigm closely and was developed specifically with the end user in mind. In fact, because of the rapid way the company was growing as well as the regional nature of its business, developing a platform that was user-centric was key to the ongoing success of the company.

If you haven’t come across Lyft, or don’t live in North America, keep in mind that Lyft is a transportation company whose mission, Sampat told us, is to improve people’s lives with the world’s best transportation system. It is currently only available in the US and Canada. Even the name Symphony, conjures up a company of many moving parts. “We decided to call the platform Symphony because if you think of an orchestration of marketing campaigns there are a lot of moving parts. Symphony’s aim is to provide the right information at the right time to the right drivers and riders so as to maximize the return on investment,” Sampat said.

Deciding on A Platform

Like many companies managing growth like Lyft, "platformization" has been a consideration as those organizations have scaled up. Sampat explained that in the case of Lyft the way the company was growing demanded the development of its own platform. “Initially it made sense for us to buy a lot of external solutions, or do a lot of things by hand,” he said. “But then as we expanded to 400 regions across the US and Canada it was not scalable to do this by hand and by marketers. So about 18 months ago we decided that we needed a platform that did this automatically.”

This included input from the tech leads who started looking at what the high level architecture should look like, and notably what key components of the platform would be: Is there an attributions layer? Is there a life-time value layers? Is there a budget layer? At this stage the project leaders started evangelizing the architecture across the company, to the leadership, and to the stakeholders to basically get buy-in. Essentially, the entire process was agile driven, Sampat added. “We definitely followed the agile development methodology in software and in processes as well,” he said.

There were two main challenges. One was the technical challenge and a challenge that had many moving parts. “There was a lot of third party data from search, data from spend aggregators, data form social channels — so the really big challenge was how to get clean data so the machine learning  can  make the rights decisions. You’ve probably heard with machine learning: garbage in, garbage out. So we wanted good data in and good data out,” Sampat said.

The second challenge was that once the marketers were freed up and no longer overloaded with day to day activities, they had to adapt their role and engage in the more creative aspects of their work.

Both challenges are being managed though unit testing, data validation and other technical applications in the case of the technology challenges while the marketing problem involved a lot of education, hand holding improving observability and showing marketers the benefits of the system. “It was very important that they opened the black box to the marketers so they could understand what was going on,” he said.

Final Thoughts

There are many ways to scale agile to multiple teams, including SAFe, LeSS and StarCIO.  According to VersionOne’s State of Agile Report, as of 2018, 97% of organizations even at that stage practiced agile in some form. Respondents reported that this adoption is not always widespread within their organizations, but it is clear that 18 months later it is spreading. As we approach the 20th anniversary of the Agile Manifesto next year, it is certain we are going to be talking a lot more about this in the coming months.