9 Keys To Building A Truly Data-Led Organization
Truly? Yes, truly. Let’s face it, most leaders of Marketing organizations – and many other types of organizations for that matter – talk a good game about the importance of being data-driven. It’s a good sound bite, in large part, because it’s both rooted in common sense and it embodies the modern in “modern marketing.”
But actually being data-driven is a very different thing. A fraction of organizations really nail it when it comes to having data be at the forefront of even most decisions. What are the keys to making the data-driven sound bite ring true? Here are nine of them…
Create an always on testing program
In digital, nearly everything can be measured and tested. That doesn’t mean everything should be tested. But it does mean that your organization has to have the capacity and resources to be testing on a continuous basis. Data-mining rovers should be bouncing across your digital landscapes 24/7.
Combine analytics with action-oriented measurement activities such as A/B and multivariate testing. Equip yourselves with the skills and people power to constantly test your most important user flows. From the checkout flow and the quoting engine, to the subscription form and your top campaign landing pages – test continuously.
Check egos at the door
Sometimes the data from these testing programs will tell us our precious creations aren’t actually all that great. Whether it’s copy we’ve written, designs we’ve crafted or a concept we’ve spent much time conceiving, we must embrace fail-fast, digital environments.
Organizations simply cannot be data-led organizations unless the prevailing culture subscribes to a data-led way of thinking. What creates these cultures starts with the individuals that comprise them. Yes, leaders have to believe in and evangelize data-driven frameworks. But individual contributors too need to be okay with having data prove them wrong from time to time.
Shift to a multiple idea mindset
If creative contributors are hesitant, one way to get them on board is to have them craft multiple variations in their initial creative process. Instead of choosing one image through guesswork, choose six and prepare them to be tested against each other. Role the dice and see what happens. Instead of crafting one headline on that landing page, come up with four from the very beginning. Watch the headline horse race unfold.
By bringing testing into the mindset this early on into the creative process, the creators can pivot from trepidation of being second guessed by data to actually having fun with data.
Don’t confuse data with Simon
When Simon says take two steps forward, we take two steps forward. Data may suggest it’s better to send visitors down path B, but that doesn’t mean path A isn’t better for business – in the end. Data is great at giving us insights into direct, “cause and effect” relationships. But we humans are better than data at sizing up more complicated nuances and dependencies that often make up the bigger picture.
It is important to understand that data-producing practices such as A/B and multivariate testing do not preclude humans from making the ultimate decision. Data can be the exclusive decision maker. This is where recommendation engines, AI and machine learning shine. There is a place for that. But when nuances and dependencies are at play, we humans can interpret the data and still make our own decisions.
The key here is that a commitment to regular testing does not take decision making powers away from us. The data should be viewed simply as a strong input into our collective decisions.
Construct an optimization layer
To be truly data-led, data-driven mentalities have to permeate the key parts that make up the whole. This doesn’t require putting every individual contributor through data-driven boot camp. But there should be an optimization red thread that cuts across functional teams.
For example, in Marketing, an optimization layer would cut across teams that design, write copy, source and create content, drive social media messaging, develop site navigation, manage paid media, map out the user journey – the list can go on and on.
Sure, testing practitioners can set up the experiments, analyze data and socialize the results. But without involvement – and some skin in the game – from the different functional groups in the optimization process, a widespread data-driven culture will be difficult to take hold.
Don’t be afraid to take risks
Testing in digital often validates great ideas. What is equally great about testing though is that it acts as a safety net for when our ideas don’t pan out as expected.
When combined with stashing the ego in the dumpster, “failing fast” means you can think BIG. Big thinking breeds success. The sheer existence of a solid testing platform should be an enabler to big thinking.
As Mike Damone once said “it’s like riding a bike. Fall off; you’re right back on.” This applies to iterative testing in the digital world. Take a hit in the form of a negative lift. Keep testing and you’ll steer your way into success, more often than not.
Have an executive sponsor that believes in data
While individual contributors must also buy into a data-driven process, there is no denying the fact that success in building a practice of anything often stems from leadership’s belief in that thing. From a purely practical standpoint, an executive sponsor is needed to foot the bill to build the core practice.
But equally important they need to drum the beat – repeatedly evangelizing the importance of producing data and letting data be a key a contributor in decision-making processes. Having leaders that are wired with this mindset can make the development of a data-driven, data-led culture, much easier.
Take action on the data
A healthy data-insight-action pipeline has a massive snowball effect on business value. It’s increasingly common to have analytics and testing practices. But what often gets lost through organizational deficiencies and competing priorities is the notion of actually doing something with the newfound data that testing is uncovering.
Pivoting from data and insights to driving widely scaled implementations often requires coordination with other groups – that manage and publish the websites, for example. Particularly if the aforementioned “Optimization Layer” is not established, competing priorities can get in the way of the data-driven action.
The distinguishing factor of a data-led Marketing organization is the existence of a healthy data-insight-action pipeline that is clog-free. One where the process of testing and optimization doesn’t stop when the tests are over.
Leverage agile development frameworks
When a program is cranking out dozens or several hundreds of experiments a year, it is hard to pivot from data and insights to implementation in a waterfall development process. When there are fewer development releases spanning a yearly cycle, the prioritization battles are likely to get in the way.
Optimization is often seen as a “nice to have” relative to core, fundamental content and functionality. So when push comes to shove, core content and functionality gets pushed and optimization gets shoved.
What can address this is a commitment to agile development processes – shorter, more nimble, more frequent bodies of work – that can enable optimization work to “hook in,” in a more seamless and sustainable fashion.
Nail these 9 keys and you’re on your data way…
Originally posted on LinkedIn