I’ve been thinking about a question posted by Rohn Jay Miller over at the Agile Marketing Facebook group. Rohn Jay asked (and answered) “what does it mean for a business process to be agile?”
This is a great question, and one that I’ve been thinking about and blogging about for some time. I’d describe a business process as Agile when it possesses five characteristics, described by the five phrases below:
Small Batches
Waterfall development methodology delivers software at infrequent intervals, in big batches. Problems arise when wrong assumptions aren’t discovered until the end of the process, and when the business requirements change faster than the software can be shipped. Agile development methodologies address this by breaking down large projects into smaller, bite-size chunks, so that software can be delivered in rapid iterations, incorrect assumptions can be discovered and fixed faster, and the business can realize value earlier. Because the amount of time between the specification of the business requirements and the delivery of working software is shorter, there is less probability that the business requirements change before the software is delivered.
Agile Marketing also works in small batches, using what Neil Perkins calls a test and learn mentality. Eric Ries, in “The Lean Startup“, calls this the build-measure-learn feedback loop. Scott Brinker, in his Agile Marketing Manifesto, suggests “Numerous small experiments over a few large bets”. The concept of small batches can be traced back to Japanese lean manufacturing, but the purpose here is different – it’s about experimentation and rapid learning. Agile processes are iterative by nature, and the iterations are in small batches.
Data Driven
Determining the outcome of a small, agile process is not a matter of opinion, but a matter of measurement. Just as agile developers write the acceptance criteria before they code, anyone designing an agile process needs to decide how to measure the success or failure of the process before the process begins, and collect the data. The data determines, for example, which of two alternatives is better in a classic A/B test. In many cases, agile processes apply the scientific method, formulating testable hypotheses, ones that can be proven right or wrong, and testing them. Particularly in marketing, the success or failure of a campaign has often been a matter of opinion, rather than a matter of measurement.
One of my favorite stories comes from “The End of Marketing as We Know It” by Sergio Zyman, the former CMO of the Coca-Cola Company. Sergio was in a meeting with Roberto Goizuetta, the CEO of the Coca-Cola company, who said that he didn’t like Sergio’s latest TV commercials. Sergio replied “Look, Roberto . . . If you’re willing to buy a hundred percent of the volume out there worldwide, then I’m happy to do advertising that you like. Otherwise, I’ve got to keep doing it for those damn customers.” Roberto got the point immediately, and they reviewed the numbers, which showed that the commercials were working.
Agile processes are data driven.
Responsive to Change
United Airlines took several weeks to acknowledge and reply to Dave Carroll’s United Breaks Guitars YouTube video, and by that time, the damage was done. The Red Cross, on the other hand, responded quickly to a recent Twitter faux pas, where an employee tweeted about Dogfish Head beers on the official Red Cross account, and as a result, the Red Cross realized a surge in donations. Agile processes must be real-time, and react quickly to changes and opportunities in the marketplace. David Meerman Scott has written extensively about the need for real-time responsiveness in this always-on, 7 by 24 world, in his books “Real-Time Marketing and PR” as well as “Newsjacking“.
Responsiveness to change also means responding to long-term trends as well. For example, Coca Cola marketing, now under the leadership of Jonathan Mildenhall, has embraced Content Marketing as a long term trend. The very nature of Agile implies flexibility, and the ability to change with changing conditions.
Experimental
Agile processes move forward by experimenting, by taking calculated risks and trying out new things, while measuring their effectiveness. In Agile, failure isn’t a sure way to get fired, but celebrated as learning. Many people forget that one of the biggest “failures” in the history of marketing, the introduction of New Coke, ended up re-vitalizing the brand and re-invigorating customers’ relationship with “Old” Coke. More recently, in the Content 2020 video referred to above, Jonathan Mildenhall suggests a 70/20/10 approach: 70% of your investment should be in what you know works, 20% should be spent on innovations that riff off of the 70%, and 10% should be spent on wacky, off the charts ideas.
Visible
Agile development processes like Scrum, Sprint Reviews and Burndown charts bring new levels of visibility to the development process. They encourage communication and transparency, so that problems get fixed quickly, without finger pointing, and the pace of development improves over time as teams measure their velocity. Every agile process can benefit from visibility. Just last week at SXSW, Todd Park, the new CTO at the White House, spoke about the adoption of lean and agile principles in government. He pointed to one early success where the government made health care data much more visible through public access and APIs. The result is hundreds of applications that use this data in new and innovative ways.
So there you have it, Rohn Jay and everyone else, my somewhat long winded answer to your question. What do you think?
Your provided knowledge is very helpful for me. In agile development life cycle the initial planning and analysis is kept to a very high level just enough to outline the scope of the development project. The whole team goes through a series of iterations, analyzing, designing, developing and testing each feature in turn within the iterations.