In my last piece, I covered off how to start laying the groundwork for new analytics projects within the business when you are moving beyond the traditional use of Business Intelligence (BI) for centralized reporting and dashboards. The goal for these projects is always to make the company perform better.
However, that idea of “better” is a potentially thorny one. Like many things in life, making data “better” can mean different things to different people, and each of these can overlap. Here, we’ll look at data quality, making data available and making it actionable.
Improving Data Quality
The foundation for successful analytics programs is data – without good data, it’s impossible to provide useful results back to the business. Making all sources of information accurate is one way that “better” data can be created.
Data quality is one discipline that is important for all companies to get right, yet few currently are. According to Experian’s Data Quality Global Report for 2015, 92 per cent of respondents worried that their data was incorrect on some level, up from 86 per cent the previous year.
To improve data quality, it’s important to look at how data is treated across the business. This includes making sure that data is being stored correctly and that the right information is being entered in the first place.
For applications that collect information from external sources – potential customers entering their details on a website, say – re-designing the data collection process can help. The aim here is to ensure that the quality of new data is better from the start.
For internal applications like Customer Relationship Management (CRM) or service desk, getting the right data entered is essential. While there are tools that can automate some of the processes around data cleansing and deduplication, it’s important to educate staff on what data quality should mean to them. Putting things into context can help this be less of an abstract request and something that matters to their performance as well.
The goal of data accuracy should be in place across all data entry/update processes but end-users can also improve data quality during the analytics process. Some business intelligence platforms have introduced data cleansing and deduplication capabilities that help end users improve the quality of data they extract from sources before they analyze it. This helps people ensure that their results are more accurate as well.
Making Data More Available
Alongside looking at each source of information on its own, it’s important to look at how those sources are combined and used. Today, many business units have to consider more than their own goals. For example, marketing teams used to be measured on the quality of their branding work and increases in recognition amongst target customers. This is no longer the case.
More and more company operations have moved over to digital processes, so it is now possible to track more activities that are carried out. Social and cloud applications assist marketers today in seeing how much interaction is taking place around their branding campaigns too.
However, linking campaigns to revenue generated or sales made is difficult as the data for this is held in different applications and systems. To build up a full picture of customer interactions means bringing together data from multiple applications across the business.
Currently, there has been a rise in the number of users for visual data discovery tools, which can help people to create their own data visualizations quickly based on combining multiple sources of data. These tools can help immensely when it comes to making data “better” for individuals.
However, this only works at the personal level. Each user combines their own sources of data, which may be stored locally on their machine or taken from reports; however, sharing the results of this analysis can be difficult. Other users may have their own data visualization or BI tools, while the data being used can also be different depending on what people are inputting in the first place.
In these circumstances, users can only conduct analytics around sources of data that they have; they may not know about or think to use data from other external sources or applications that are generating useful material for analysis. This means that it also can be difficult to build up and maintain a consistent view of the business. Without centralized mechanisms to ensure consistency, the growth in available data sources increases the risk of conflicting interpretations of the data between people. Consequently, this can lead to a lack of confidence in decisions made.
The dilemma here is that traditional centralized systems can be too unwieldy to deploy agile BI projects, while the data discovery tools don’t offer that ability to join up data across the whole business in a trustworthy way.
Instead, it’s worth looking at how the data governance can be centralized while the access side can be opened up to more users, making it available to them in new ways that will help their decision-making get better over time. While the management and interaction between data sources can and should be governed centrally, the user layer should instead be opened up to users for them to conduct their own queries and analysis.
A successful approach to making data more available enables end-users to access and combine data in their own ways. IT can provide guide-rails to make sure that data is governed consistently and used in the right way, but does not have to be prescriptive in terms of the reports created or questions asked.
Empowering users in this manner, with the proper controls, can help with building greater links across the business. Rather than looking at only marketing data or sales data, a fuller picture of the customer can be put together based on operational and support records as well. This can then be shared across the business.
Making Data Actionable
This increase in the availability of data is not the end of trying to make data “better” for the business. Instead, it provokes people to think about how the results can be used over time to create new ways to measure how the company is succeeding.
A good example here is to return back to the aims of the marketing team. Rather than looking at social activity or even web-site traffic, marketers can link customers back to specific campaign steps and produce a guide on the revenue that was achieved. By providing this firm link between creative campaigns and revenue, Chief Marketing Officers can demonstrate where they provide a wider return back to the business.
Analytics here can provide results on how customers responded to different touch-points across multiple campaigns as part of their buying journey, or look at the big picture return from marketing budgets; what is important is that this real-world insight is available to be used. More importantly, CMOs can use this data for faster insight into campaigns that are producing results right now, and which should not have more investment put into them.
For other management leaders within the business, this ability to look at the whole process is essential for their roles as well. Supply chain managers can see where they are spending budgets, but more importantly spot where there are more opportunities to improve turn-around times and efficiency. On the sales side, analytics can tell sales people where they can look to sell specific products for the best chances of success with given customer segments.
By making more use of data from across the business, it is then possible to look at how to redesign the workflow and process that people are working to. This can go hand-in-hand with rethinking the metrics that are used to measure success over time. The aim here is to provide more tangible evidence that each decision has a positive impact on the business.