Maybe you realized that you need a better sense of where you are financially in your business. Perhaps you want to understand whether a new opportunity would be a good risk to take. Or perhaps you are one of the very few starting your business with analytics built in from the beginning.
However you arrived, you’re ready to take that first step into intentional data for your business. Here are seven key questions you should answer as you get started.
1. What is driving the desire to bring in more analytics?
You likely arrived at analytics because you are trying to improve a single area of your work, your team, or your organization, or there’s a specific challenge that arrived that you need to solve. For example, maybe you want to do a more effective job selecting your non-profit’s programming by informing that choice with results or outcomes data. Perhaps you’re seeking improved customer retention, or need to rein in increasing expenses in your manufacturing. Whatever the reason or reasons, you need to be able to clearly articulate the why for your analytic shift. The “why” will help you determine what, out of the infinite world of possible data, you want to focus on first. Data analytics always starts with a question, and your challenge or opportunity gives you your first question(s).
Also, having a meaningful and inspiring why will help you and your team make it through the challenges and pain of change. If you aren’t already using data in this area, the will be change, and change can be uncomfortable. A good “why” keeps your eyes on the prize.
2. What is the scope of this “why”?
Will you start with a single project, several related projects, or are you going whole hog with a full organization culture change? Whatever the scale, you need to define what success will look like. Perhaps it’s a customer retention rate of 30%, or a 5% reduction in material expenses. Or perhaps it’s that your CEO receives and uses a daily dashboard in his decision making.
You need to know this scope and what success specifically looks like, because it will lead you to what data you need to track for for the goal and for the analytic project itself. You have to drink your own Champaign – who will want to do the work to bring analytics to bear on a challenge if the analytic project itself isn’t following its own advice?
Especially if this is your or your organization’s first analytic endeavor, consider breaking it into phases if it incorporates more than one discrete project. Pick the one that will have a measurable impact, both in terms of size of the impact and actual ability to quantify the impact.
3. What data elements and analytic processes will you need to achieve success in the project(s) you defined for question #2?
You’ll need two sets of metrics. The first are the ones that you need for the actual solution or change you’re trying to achieve. In our expense example, this might be tracking the components that make up total expenses and the percent change over time to identify the major drivers of expense and track progress in reduction.
Then you’ll also want data on the analytic process itself. How much does your new system help, and how much in time or money did it cost?
4. What is the current process, and how well does it work?
Your project is only a success if it works better than what you had before. How is your CEO currently making decisions? How is the supply line currently managing expenses? Even when something’s not working well, it’s probably working well enough or people would be doing something different. What can you learn from the current process? Where can you provide improvement with analytics?
You may find that you want to adjust the metrics you designed in step 3, perhaps based on what you and your team can actually control, or progress that’s already being made. This is also a good point to consider what challenges there might be to the coming changes. If you’re going to add steps to a process, for example to collect additional data, how much time does that add to the process? Can the process handle that addition time, or will something else need to go?
5. What data and analytic resources do you have?
Now that you know what you are trying to accomplish and where you’re starting from, it’s time to take stock of what you have on hand to help you succeed. This includes inventory of all the obvious data resources and tools – BI software, analytic staff in your organization, existing databases or datasets.
But it should also include a review of the less obvious things that might not seem like assets at first glance. Is there someone who’s job title doesn’t involve analytics at all, but has some coding or statistic skills? Is there a process that is generating important information for your project that’s not quite being captured yet? For example, maybe you find out that the membership intake form for new customers asks a few questions. This might be a great place to add a critical survey question about retention.
The primary goal is to secure everything needed to produce the metrics you outlined to answer question #3, but it can also help to know what else is out there in case it opens up new questions or opportunities.
6. What infrastructure and people do you need to create and sustainably manage your new analytics?
Even if this first take is a small project, don’t do it alone if you can avoid it. Analytic champions are much more like the Avengers than Chuck Norris, with each team member providing unique superpowers rather than one Lone Wolf doing everything. More eyeballs means less chance for errors to get through, and more creativity in designing the analytic solution.
Plus, you absolutely want to include the end users and front-line workers with the business knowledge needed to inform the context of your analytics, as much as you need technical support with the data.
Don’t be afraid of a bit of manual work to start out with, especially if it keeps the project within the current skillset of you or your team. Start small, and you can build on your success as it happens. It’s hard to know for sure what software, databases, etc. you need until you really get a sense of your data, your processes, and what works best for you/your team. While you don’t want to let a suboptimal system get too entrenched, and you certainly want to be aware of how much technical debt you’re accruing, it’s often best to hold off buying anything new until you have a defined and sustained need. You may find ways to adapt your current system, or if you do need to bring in new software or switch to a new vendor, you can phase in those changes once you really know what you need.
7. How can you make sure you’re improving and course correcting both your intervention and your analytic processes?
Discuss with your stakeholders what a good Ideate –> Implement –> Assess –> Iterate process looks like to them. Build into every data project the expectation that you will monitor success and adjust as needed. Have a change management plan from the beginning that includes established benchmarks for your progress metrics that will trigger decision points/actions to try something else.
This is true for both for your goal and for the data process itself. For example, you may say that if customer turnover goes above X or if outcome for a project is below Y then you will try intervention Z instead of A. For your data process itself, you may set that if manual time required goes above X hours or if error rate goes above Y, or lag time is greater than Z, then you’ll move to a new software or bring in outside support.
Congratulations on taking this critical step towards data-driven decision making.
