Focus on People & Process Keys to Better Decision-Making Ability
What Analysts Want
In the song “I.G.Y.”, Donald Fagan reflects on the heady optimism of the world in 1958. He describes a technology-enriched future, where even making decisions is automated:
This dream’s in sight, you’ve got to admit it…the future looks bright…
A just machine to make big decisions, programmed by fellows with compassion and vision…
What a beautiful world this will be, what a glorious time to be free
Today, automated decision-making may increasingly be used in various business domains (Davenport, 2005). Typically, however, digital analytics teams grind it out – they collect and scrub data, select, learn and use tools, create reports, dashboards and analyses, run in-market tests and, hopefully, get some sleep in between. But being perceived by clientele as adding value and “coming through” with data for key decisions inspires most.
In decision-making it all comes together—the people, the tools and the data—with great opportunity. There may be countless times where, in addition to revenue, expenses, headcount, competition, etc., customer behaviors and sentiment collected from digital channels can inform decisions. Robust alignment to decision-making can drive evolution of underlying capabilities for data management, reporting, analytics and technologies. And better capabilities enable better, faster decision-making. Getting it right creates a “spiral of success” (what a beautiful world this will be!) But the flip side is a disastrous wall of misunderstanding between analysts and their stakeholders.
So how can we support an unquenchable thirst for data and, ideally, take advantage of opportunities to support the business? How to avoid losing the value of digital analytics, and have the right data to inform decisions available at the right time, right place and in the right formats for the audiences that need it?
There is no magic answer and indeed it can feel overwhelming, but there are two key points:
- Technology can help, but there is no substitute for understanding the subject matter (marketing, customer service, product distribution, etc.), the subject matter experts, and how they think.
- It is important to keep in mind the adage “perfect is the enemy of good”. Aiming to continually improve the decision-making ability of the organization should be the goal, not to support every possible decision.
Start With the Process
Prepare a map that reflects the unique process stages for your organization. There is no need for highly detailed documentation but, for each stage, start by listing key inputs and outputs and the sources and consumers of each.
Just mapping the process can lead to useful questions. What decisions are made during each stage? Are there formal go/no go “gates” between each stage and, if so, how are these decisions made and by whom?
Multichannel marketing is used as the leading example (see diagram), but the same concept applies to customer service, e‑commerce, etc.
At each stage, from planning through closing, there are needs and opportunities for analytics – from designing tests to ensuring data are properly collected to report distribution. So at each stage, there are hundreds of questions one can ask to understand decisions made and generate ideas for improvement:
How much will be spent on each channel in the campaign and why? Why was this audience segment chosen and not others? What does success look like for each audience? How will the site / email / banner ad / fan page / PPC elements etc. be designed and why? If there are not multiple designs for each tactic (to enable marketplace tests), then why not? How are media formats and languages (Flash, JavaScript, AJAX, etc.) chosen? What assumptions are being used for “the fold”? What data need to be collected and managed? How are data to be collected? Is there a repository of brand or campaign information that can house (or link to) reports and analyses you create?
Using a methodology like Lean Six Sigma (LSS) is highly recommended for analyzing processes; it’s a robust and proven approach that can improve abilities to support decisions (Hamel, 2012). But as Hall of Fame pitcher Steve Carlton—known for minimal time between pitches—would say “if you think it long, you think it wrong”. If you are unfamiliar with LSS, get started informally and take small steps. But get started.
Maintain this information as a living document, creating an integrated view of marketing and analytics. Link it to existing analytics documentation such as for data collection or “tagging”, testing methods, tool manuals, etc.
Determine Who’s Who and Walk in Their Shoes
For each process stage, identify decisions-makers and key players, both business and technical. Build relationships, but don’t just interview stakeholders; join them as they go about their business. Attend formal decision-making meetings such as budget allocation. But also invite yourself to concept and design reviews and presentations on market research results. Listen to the questions that team members ask. Seek first to understand.
Don’t try the direct approach. Analysts sometimes ask marketers “what decision will you make with these data” in response to a request. It’s often impossible for line managers to answer such a direct question. We all make decisions, consciously or unconsciously, all day long: How much coffee should I have? Should I turn at 15th or 16th street? Do I need to fill up the tank? What time do I need to leave work to pick up the kids?
Where do you sit in relation to your clientele? In a room down the hall, the next building, perhaps in another state or country, connected by technology? Co-location eliminates barriers, and you can “feel” and “smell” how marketing works, and understand how marketers think. This can’t always work, so find other ways to be a part of the team: If you normally attend meetings, stop by when there is no agenda. If you tele-connect from another site, show up in person now and then, or use video occasionally. Personal connections mean everything.
Try Something Different
You have mapped out the process, better understand your stakeholders, and have several ideas to improve your service. Now is the time to try something different. You may need to start small and then try something requiring more effort. Only you can identify what may improve your unique circumstances, but if you are stumped, here are some thought starters:
- From an analytics viewpoint, decision-making depends on the elements of data, reporting, analytical methods and even how we define things (not just metric definitions like KPIs, but also “business” definitions, such as “what is a customer?”). If you are not sure where to begin, drill down in these four areas.
- Combining “clickstream” data with customer, product, or geographic attributes almost always improves decision-making because it creates new ways to segment customer behaviors. If your marketing program segments customers by demographics, add census data. If psychographics is important, perhaps product features can help.
- Consider more “data democratization”, make data as freely available to as many stakeholders as possible. It can virtually eliminate the cycle time for getting data to the intended audiences. But “raw” data in the hands of its consumers only goes so far; reports without context are nearly meaningless. Excellence in data management, stakeholder education and using social media for report distribution are some keys to success.
Look at the Bigger Picture and Repeat
You’ve implemented some improvements and stakeholders love it.
First of all, make sure your boss knows— ideally have your stakeholders tell your boss instead of you. Use the opportunity to get the support resources, etc. for the infrastructure improvements that would make a difference. (“You know, if we used tag management, we could spend less time cleaning data and more time helping So-and-So on the brand team”).
Then it’s time to “wash, rinse, repeat”: when the time is right, look at an even bigger picture. Each brand or service may have multiple campaigns, and in turn, enterprises or business units have many brands or services. Decisions need to be made, and lessons learned, across campaigns and across brands or services.
Do the same as above: identify players, build relationships, and ask critical questions. How do brands (or services or franchises or business units) prioritize campaigns? How do enterprises allocate budget across brands? What are the common themes between brands? Is the definition of success the same, or at least aligned, from level to level?
In Summary
Decision-making is a complex and often unconscious act for marketers and line managers. Improving analytical services and capabilities for better, faster decision-making requires a deft combination of art and science. Technology can help, but there is no substitute for a deep understanding of the process at hand and the people involved. Creating and using a process map and cozying up to marketers and other subject matter experts can shine a light, for digital analytics teams, on the path to adding more value.
And if you need a way to break the ice with your clientele, buy spandex jackets, one for everyone (Fagan, 1982).
©Joseph Reach, 2013
_________________________________
Davenport, T. H. (2005, July 15). Automated Decision Making Comes of Age. Retrieved December 2012, from MIT Sloan Management Review: http://sloanreview.mit.edu/the-magazine/2005-summer/46414/automated-decision-making-comes-of-age/
Fagan, D. (Composer). (1982). I.G.Y. [D. Fagan, Performer] On The Nightfly. http://donaldfagen.com/disc_nightfly.php
Hamel, S. (2012, November). Lean Six Sigma for the Digital Analyst. Retrieved December 2012, from Online Behavior: http://online-behavior.com/analytics/lean-six-sigma