When rolling out analytics services to teams within a business, it’s important to bear in mind that there is no “one size fits all” approach that can work for everyone.
In the past, traditional BI implementations were limited to specific reports that were created and used by analysts for the management team, while any ad hoc report requests were filtered through these specialists, as well. The users involved in BI were small in number and knew exactly what they wanted to produce.
Contrast this with today. Business units want to use data for their own decision-making purposes and won’t put up with delays in getting access to the reports that they use for these decisions. If central BI or analytics teams can’t or won’t support them, then they can and will acquire their own tools for analysing data. According to the firm’s Magic Quadrant report for Business Intelligence and Analytics, Gartner sees this changing the BI market landscape and has assumed that most business users will adopt self-service analytics tools that can prepare and visualise data by 2018.
This shift opens up data for individuals, but it also represents a dilemma for central IT. Should we let each department run with their own solutions or look to take care of their requirements like before, only faster? From a data governance perspective, letting self-service take place without guide rails in place on what and how data is used can lead to potential problems.
Instead, it is worth looking at how to support line of business teams in getting self-service access to data while still supporting central data management. To achieve this means looking at how each team will use data over time and where the same data will get re-used for different purposes.
Thinking about data use across teams
Businesses exist to sell products, while organisations in the public sector are focused on serving citizens’ particular needs. To achieve these objectives, organisations typically rely on line-of-business teams to extract insights from data in functional areas such as sales, finance, marketing, procurement and the like. This becomes a potential problem for analytics when each team has its own targets for success in place.
Solving this requires smarter use of data among all the teams involved. This does not mean prescribing reports to be used or sticking with existing BI approaches. Instead, the BI team can look at how each team uses data to achieve results.
Getting started involves looking at how each team currently defines success. For example, the marketing team may look at the volume of leads created and passed to sales as its overall target, with further analysis possible on the number of leads accepted by the sales team after qualification. Conversely, the sales team may look at the volume of appointments created and the number of conversions to new business, regardless of the source of those leads.
In this example, both the marketing and sales teams care about net new business opportunities closed. However, there is a potential gap that exists between the quality of leads created, compared to closed business. Each team can point to its own results independently, rather than looking at the whole lead-to-revenue process.
For IT teams looking to support greater use of analytics, getting a focus on the business process can ensure that all teams are becoming successful in their use of data. Each team may continue using their own measurements of success, but they should all have access to one trusted and consistent view of the business, even if data comes from different sources.
Achieving this will involve networking data sources to provide more context for analysis. In our example, Marketing applications such as web analytics, marketing automation and CRM data can provide information alongside the sales force automation and finance systems. Based on this, marketing can then look at how many leads were created and passed on to sales, while sales can start looking at which customers are more likely to convert and how much business is created.
This inevitably involves distinct dashboard and visualisations for each department. However, by looking at this information in context, there is more opportunity for collaboration and discussion, between marketing and sales leadership, about prospect/customer targeting. Information-based decision making can lead to smarter use of budget by marketing when targeting prospective customers, while sales can get a better overview of which target customers are more likely to close, based on criteria such as market size and vertical industry.
Serving data for self-service teams
Some teams may have brought in their own desktop data discovery tools already. If this is the case, they may have some experience with data visualisation and using data for decision-making. However, this will normally focus on the team’s or individual’s own objectives rather than looking at this data in a broader context. To help these teams move from looking at data in a silo, it’s worth going through the following areas:
- How does the business process create data?
Thinking big is not just something for the CEO. It is something that everyone across the business can get involved in. However, this can require some changes to how data is used over time. Rather than looking at departmental goals as the final reporting metric, this can be an opportunity for each person to think about their individual role in helping achieve the bigger company goal.
To start with, existing metrics and reporting can be used to define success. In marketing, this might be looking at leads generated through social media campaigns or traffic to the website. However, over time, this can evolve to look at the entire customer journey and how the team moves customers from initial prospect to conversion. This can then lead to changes in metrics based on what the business needs.
- Are the metrics right?
Once these wider business processes are described, it’s necessary to identify the data sources that support these processes. Many companies are going through digital transformation processes, which can generate more data for companies to track and measure. However, these metrics may be based on data that is then split across the business.
As companies move to digital initiatives, changing these metrics can be a big challenge. This involves understanding what differences are involved over time, and which teams will provide information to complete these metrics.
As an example, companies running their businesses on a “software as a service” (SaaS) model, rather than selling individual products, may be looking at how they can measure their success. Traditional profitability models are not as useful for companies that base their operating models on ongoing revenue, so looking at profitability over time is much more useful as a measure of success. However, this involves getting much more detail on which customers sign up for long-term contracts and cancellation rates, rather than just the acquisition cost involved.
By analysing how long a customer remains on the service, compared to the average cost to win that business, companies can get an overview of Customer Life-Time Value, or CLTV. CLTV can then be used to see where marketing and acquisition investment can produce more profitable customers and increase investment there. Conversely, if a segment of less profitable customers can be discovered, this group can be reduced or avoided.
Making this shift involves data collaboration between business units, as well as understanding the impact that each team’s decisions can have on the overall business.
- How is data governed over time?
Alongside this change in metrics, it is worth looking at how and when the data created is managed over time, as well. Does this data get updated in real time, or is it created on a regular basis such as every week or month? Is this data managed by a central IT organisation, or is it “edge” data generated and/or managed by end users?
This can have a big impact on the teams that use it across the organisation. For teams that can effectively work in real time – marketing and sales are prime examples – data can be taken and used on an ad hoc basis. For Finance – where billing values might be created once a month, every month – comparing sales figures at different points in the month might create different metrics and, therefore, affect decisions in different ways. Instead, it’s important to provide more governance and control over this data so that everyone is looking at the same data values over time.
This can be achieved through a networked BI model that connects different data sources and presents them as one, empowering teams and individuals to leverage and extend the network. By bringing these BI data sources together into a network, each team or individual can ask their own questions and get answers based on the same data.
While the results of that analytic work might be different for each team, the data underneath would be the same.
Pedro Arellano, senior director of product strategy, Birst
Image credit: Shutterstock/Sergey Nivens