What’s for dinner? It’s a common question everyone in your household asks regularly, and maybe even more often this past year while spending more time at home. You will enjoy cooking a meal with the family more when you consider first your utensils, your space and the right ingredients. You want to think about these basics before you jump into lavish appliances like the air fryer and food processor and advanced techniques such as making pasta from scratch.
Considering basics first is also a good way to approach an enterprise analytics strategy. Before you jump into machine learning and artificial intelligence applications, you need to consider data as your key ingredient, the core of your analytics capabilities. Without basic data management and unification capabilities, your analytics engine and prowess will only take you so far.
Figure 1 – An analytics strategy is optimally built progressing through these capabilities
You could start at the top layer and produce analytics outcomes immediately in the form of dashboards, machine learning models and unstructured text learnings. But pretty soon you would see there are common activities you address each time you face the data set for a new analytic question. These rote tasks cost you time and money, and it’s time to think about your data and your processes differently.
The bottom layer of Figure 1 includes fundamental capabilities that inventory the data you have in-house, raise the quality of your data and enable access to it. You cannot build good analytic outcomes from poor data. You need global data management practices across systems, functions and geographies.
- Collect data and inventory its source and the date-timestamp. Address questions such as, what are you communicating to end users about how you will use the data? How will data collected help you better understand people or drive new business opportunities? What other data sets combined with your data will bring you richer insights?
- A data storage plan is critical for a data architecture allowing shared accessibility of company knowledge. You will want to consider how often you need to access data and the total costs of data storage.
- Bringing together and assembling data through processes such as quality control, quality cleansing, and de-duplication make up this housekeeping practice. The goal of assembly is a robust set of processes and rules to drive high-quality, consistent data into the organization.
These are basic tools that when used repeatedly will offer you faster, consistent performance to produce analytics outcomes. Just as you prepare a meal with your family, you want to know which tools and ingredients you have, where they are located and their quality or shelf life for use in the meal.
Strategic Takeaway: As part of your analytics strategy, know what data exists and assess data accuracy and completeness. We are seeing many consumer packaged goods companies tackle the challenge by collecting first-party data and documenting data inventories. For another industry, such as the pharmaceutical sector, new data sets may need to be acquired to tackle rare diseases.
Once you know the data you have at your disposal and are creating more complete data sets, you will want to layer more capabilities that build on your global data management practices. The middle level is what Acxiom calls data unification capabilities. Here you start to break down siloed data sets and prepare for enterprise-wide data sharing.
- Through the creation of a common identifier, identity methodologies enable understanding the omnichannel touches a person has with your brand. You want the ability to stitch together a person’s awareness, consideration and purchases with your brand across your product and customer service silos.
- It is imperative for organizations to have solid data governance policies. Governance encompasses access to data, data correction logic, data naming standards, and data stewardship processes. You should consider the ethical use of data as a differentiator and protector of customer interests.
- Integrate technology and data architecture across the enterprise. Opportunities exist to map like fields from a business standpoint that allow you to follow customers’ journeys with your products. Does your organization connect information between channels from anonymous records to known individuals?
Giving attention to this middle layer further enhances your analytic outcomes. At this layer, you integrate your data across silos. As you are making dinner with your family, you learn how to balance different tasks across family members. You integrate the ingredients to cook the meal.
Strategic Takeaway: Introducing identity concepts for a siloed organization is critical to migrating from a product-focused approach to a customer-focused approach. The automotive sector has had time to create data management practices and is now looking across its organizations for the unification of customer journeys across dealerships, loan arrangements and service centers and integrating other product lines in its portfolio. The travel industry is planning customer experiences through pre-travel, during the excursion, and post-travel.
Finally it’s time to put your data into interactive dashboards, create self-service analytics, and apply the data to test and learn. Here you explore the data for patterns; you can incorporate more than 1,500 Acxiom InfoBase attributes into your predictive models. You can venture into such things as unstructured text for natural language processing models to discern your customers’ attitudes.
The Analytics capabilities layer includes the techniques used to find insights in your data sets.
- Measurement and reporting enabling your business practitioners to regularly consume data are great first steps for creating a data-driven culture. People become better at reading and interpreting data the more often they practice. Watching how one’s work affects key performance indicators (KPIs) instills a feeling of making an impact. Through regular report-reading of observations of campaigns, customer interactions, and sales, you instill data literacy.
- With big data and more computing power, machine learning and modeling help look at larger data sets faster, in theory providing more accurate outcomes because there are more data points from which to draw conclusions. Statistical models and practices remain at the root of this power. The process is centered around drawing lines or curves that flow through the largest number of data points on a graph to make a model. You will rely on skill sets that are rooted in statistics, computer science and data science.
- Advanced data sets and processing introduce new tools and the ability to share more data sets, enabling you to look at your products and services with an eye toward new opportunities. Unstructured text and Internet of Things (IoT) data promises to alter the competitive advantages companies find in their data.
Strategic Takeaway: Building analytics capabilities is predicated on how well you treat your data as an asset. Pay attention to the processes that help you manage data as part of your analytics strategy. The financial sector, driven by legislation, leads in developing data management and data unification practices. In turn, these companies focus on developing their analytics muscle using the tech companies as leaders to emulate. This is the time to heed Stephen Covey’s first habit of being proactive. He states in his bestseller The Seven Habits of Highly Effective People, “If you wait to be acted upon, you will be acted upon. And growth and opportunity consequences attend either road.”
The same is true for analytics; if you create solid data management and unification practices, you will get extended, repeatable value from your data. You will be able to run models, and explore the data continuously, asking different business questions as you learn more about your business. You will not have to start every project as a new one because you will already have clean, unified data.
And for family dinners, you know what ingredients and tools you have to work with. You have steps each family member can follow. The well-planned, basics-first, collaborative approach will help you whip out great meals, expand your culinary horizons, and create family memories that will last much longer than the meals you share together.
Adapting these principles to your data strategy enables your analytics capabilities to soar. You will find new avenues to explore, leveraging your data as a corporate asset.
