While ‘data is the new oil’ has become something of a cliché, it’s undeniably true. Gartner reports that 90 percent of corporate strategies will explicitly class information as a critical enterprise asset by the end of 2022. Yet, information can only be powerful when it is put to work. Which is why the UK government and industry regulators (such as Ofwat) have set out strategies that encourage organizations to put data to use.
However, making data usable isn’t always straightforward. Data doesn’t come pre-packaged into customer next best actions and strategic tips. In fact, data analysis is often compared to panning for gold. Not all data is useful and you need to connect the dots to make it relevant — when you consider the volume of data managed by the average company has grown to 162.9TB, that’s one big river to pan.
In its raw, unstructured form, data can take time and energy to collect, cleanse, harmonize and contextualize. While the potential for generating insights using analytics, AI and machine learning is huge, breaking down organizational data siloes and extracting data from complex hybrid infrastructures can be prohibitive. These challenges are not just technical either. Ensuring data is aligned for it’s intended uses and delivered to the correct audience in the correct way traverses technical and business disciplines. Ultimately, if data is not handled in the right way, the quality of the results suffer. Receiving uncleansed or even accurate data that is misaligned to its intended use can render insights unreliable or unusable — ‘garbage in equals garbage out’.
All of this may leave businesses feeling like they can’t see the wood for the trees. But, once a business gets control of its data, and understands that different use cases have different data requirements, the possibilities for making data driven decisions are endless. By taking things one step at a time, the task will be less daunting. These three steps provide a helping starting point.
Step One — know what you’re setting out to achieve
You cannot expect data to provide answers if you don’t know what the question is. Businesses must set out with a clear purpose, thinking about problems they are facing, then forming a use case that could solve them, so they set out looking for a targeted outcome. For example, an energy supplier might want to model the financial impact of the price cap — looking at customers’ ability to pay, increased service cost, and potentially a rise in bad debt. Equally, a telecoms provider may want to know where best to target a new handset or bundle offering, or to understand who best to offer discounts by understanding the profile of their most valuable customers. By identifying the ultimate goal in the context of the business value it could deliver, it makes it easier to know what data feeds will be needed to build those models as well as the types of insights required.
Step Two — cleanse and enrich your data
Next, businesses must create a firm foundation to build on, by cleansing their data to establish existing data quality. For example, looking for instances of duplicate accounts and merged records, identifying poor formatting and structure of their data, identifying invalid names and email addresses or any gaps in customer records, and so on. This is no easy task when applied to a large organization where data is collected on millions of customers across hundreds of departments and systems. However, by doing the hard work, organizations can start to understand the ‘full picture’ of each of their customers. This includes linking data across internal departments to create an invaluable single view of the customer – improving everything from billing accuracy and collections to customer care and marketing. Data can be further enriched using external data points to fill any gaps. For example, leveraging credit bureau data to identify occupiers of properties to ensure bills are correctly sent. Once cleansed, organizations can implement a policy to ensure that data is collected in the right way moving forward, from file types and archiving policies (e.g., ‘all notes must be transferred to CRM’) to formatting (e.g., ‘first name, last name’ for customers).
Step Three — applying an analytics layer
Harmonizing data through cleansing and enriching it will increase accuracy and understanding, but on its own will not drive the insights needed to build business change. The final step is to extract the insights. Most cloud native technologies are API based, helping to make data extraction timely and accessible. However, most established businesses still have several legacy technologies that are more difficult to integrate, resulting in many data siloes. An analytics layer that sits across the entire business, is needed, which takes multiple feeds into one central engine, so organizations can start mining that data to get the answers they need. This can be further enhanced by applying AI and machine learning, which can help to detect anomalies, patterns, and forecasts using hundreds of parameters that would be impossible to leverage using normal analytics tools. Data modeling should be used alongside the harmonization activities to ensure data is available and correct for its specific use cases and end-users. By applying meaning to the data it is optimized to generate the required insight. This insight drives intelligent decision making. For example, enabling organizations to understand who their most valuable customers are, which marketing campaigns drive the best long-term ROI, what channels are the most effective for customer communication, and what events may impact supply chain — the list is endless.
It’s time to take action
Acting on data won’t exactly be a leap of faith if businesses lay the groundwork, decide what they want to achieve, cleanse and enrich data, and apply an analytics layer to dig deeper. At a time when countless industries face huge challenges and disruption, data driven decision making has never been more mission-critical. As the rallying cries from the government and industry bodies get louder, increasing numbers of businesses will be putting their data to use, so it’s time to start drawing insights and taking action, or risk falling behind.
Image credit: realinemedia/depositphotos.com
Malka Townshend is COO of Sagacity.