Self-service business intelligence (SSBI) is an approach to data analytics that opens the doors for analysis of queries on corporate datasets beyond BI/Insight Analysts. SSBI empowers a broader range of business users, giving them access to analysis without the need for specialist data expertise. The result is streamlined processes, flexible enough for business users to experiment with data, enabling them to make informed business decisions independently of the BI function. Ultimately this absolves business intelligence specialists of the responsibility of generating large volumes of reports on an ad-hoc basis, freeing up time for them to focus on more strategic projects.
A recent report from Research and Markets has identified a powerful upward trajectory in the self-service BI market, with global growth set to rise from $3.61bn (US) in 2016 to $7bn in 2021. With rabid demand for more data professionals to fill roles as the influx of ‘big’ data continues to balloon across all industries, businesses are recognising the need to decentralise their data operations wherever possible.
Despite the demand and promising growth in SSBI uptake, however, it’s important to pause and weigh up the relative pros and cons of SSBI adoption for your business. It would, of course, be foolhardy to overhaul BI processes without a thorough understanding of cost/benefit and plan for implementation.
The Bare Facts
BI software licences and hardware costs alone are reportedly double those of SSBI. Internal BI implementation and maintenance account for nearly 70% of the total value of traditional BI. Besides, studies find that SSBI users typically spend 41% less on IT support per BI user – amounting to a 60% reduction in support cost per user.
The productivity drain on BI experts and the IT department in general as a result of centralised BI access is a crucial factor in many companies’ decision to turn to self-service models. Business intelligence experts have complex and challenging data to process; the generation of simple reports (for example) at short notice is detrimental to the positive functioning of this vital department. Where productivity is not operating at an optimal level, there is an apparent financial leakage.
But by decentralising, you are moving the responsibility from BI professionals to business users, whose knowledge and understanding of data science pales in comparison. This leap requires strong faith in business users to actively involve themselves in understanding the new SSBI platform and leaving it in their hands to configure the platform to their own needs and use it appropriately. Will you merely be shifting the productivity drain from the IT department to business users? Can you rely on business users not to continue leaning on IT for support with using the system following an initial training period?
The answer depends, at least partially, on the ease-of-use of the SSBI platform implemented. Unlike traditional business intelligence (BI), which requires high-level expertise in SQL and proprietary language querying, SSBI usually requires no coding or SQL skills. The dream of self-service business intelligence is democratic involvement in data analytics. Real-time data becomes accessible to all in a way that is simpler for the non-technically minded to understand and use.
To ensure that this dream bears fruit within your company is first to assess the culture, skill level, and time-constraints of your business users. Will they be receptive to the change? Are they sufficiently equipped, regarding both time and aptitude, to learn and use the new platform on a daily basis?
A 2017 survey conducted by Yellowfin looked at 33 companies where SSBI has already been implemented. Strikingly, in the majority of companies only up to 20% of employees had access to the SSBI platform at all, and of that number, only up to 20% used the system. What these results demonstrate, however, is not an inherent problem with the notion of decentralisation of BI, but a shortfall by the companies to properly plan, train, and enforce SSBI use across the company.
Another reason that uses is so limited within Yellowfin’s, admittedly small, survey group may be down to their choice of SSBI system. One of the enduring reasons businesses stick to traditional BI is that, because specialists run them, they can be tailored over time to meet individual business needs on a deeply ‘bespoke’ level.
That’s why, when choosing an SSBI platform to use, it’s essential to assess its suitability for your company’s specific needs. It must prove demonstrably improved flexibility compared with your current BI system. How easy is it to integrate the self-service offering with existing systems? Pre-packaged SSBI systems, though ostensibly good value, can fail to provide the ‘bespoke’ service necessary to justify themselves. Choose a system that offers interactive dashboards and ad hoc reporting, at least, and have in mind a thorough list of requirements before going ahead. On this point, it may be prudent to request a rundown from your BI experts themselves, and indeed to actively involve staff in the process.
Yellowfin’s survey identified the key challenge in SSBI adoption as ‘data quality’. 71% of respondents were concerned that untrained, inexperienced business users could not build accurate reports, know which datasets to use, and how to use them to answer their questions. These concerns are far from unfounded. Gartner has reported that through 2017, less than 40% of Self-Service Business Intelligence initiatives were likely to have been governed sufficiently to prevent inconsistencies that adversely affect the business.
Again, you must ask yourself how competent you expect your staff to be at handling data, and how in-depth the training requirements are for a business user to achieve data quality comparable to a data professional. Is the system intuitive enough for inexperienced users?
Open Access and Cost Saving
One of the primary drivers for companies to make a move to self-service business intelligence is, understandably, cost saving. In this area, indeed, it can be a boon. Nonetheless, heed the stark warning that this factor should come second to that of expanding access to information. The more your staff know, the more they, and your company, will grow. While the aim is not to turn business users into business intelligence experts, expanding their skill set and what they can do with the data with which they are presented has the potential to generate unique, in-depth insights and highlight connections between data points that may not have otherwise surfaced.
The alleviation of responsibility of comparatively basic tasks from your data professionals will likewise pay dividends in this area: with more time to focus on deep-level data analysis, these experts will be free to deliver the powerful insight they are trained for.
Case Study: United Utilities
United Utilities, a company which manages electricity distribution and water works in the North West of England, is an example of SSBI success in action. Nearly half of United Utilities’ annual revenue is spent on the purchase of materials and equipment (everything from sewage tunnels to office stationery). Understanding, therefore, of company expenditure across 760 contracts, 6,500 vendors, 50,000 orders, and 140,000 invoices is a complex task.
In the past, UU used external data analysis consultants, whose role involved pulling data from disparate systems and manually consolidating it into spreadsheets. The process was time-consuming and cost UU around £350,000 a year. Over seven months, staff from IT, the supply chain, and in the project and support departments (a total of 64 staff) were trained on a new SSBI system while all the company’s data was cleaned and consolidated into the new system.
Switching to SSBI has saved United Utilities £350,000 a year directly through doing away with their external consultancy. The staff now have quick and unobstructed access to spend and performance data, and the supply chain team can now focus on the strategic issues that are central to their roles. Value to the business is substantially improved.
Does SSBI Pave The Way For Self-Service Data Science?
In some ways, argues David Rostcheck, Senior Application Architect at Toyota North America, it’s possible to define data science as an evolution of business intelligence:
“Both roles are information-centric and analytical. They both make use of statistical methods to extract insight from numbers. And both call on the skills of visualisation to find new realisations and to present results in a way that others can efficiently absorb.”
However, data science focuses more on predictive analytics rather than extrapolating information on events that have already occurred (as does BI). “To investigate the unknown,” Rostcheck writes. “Data Scientists conduct experiments and form hypotheses.”
Does this mean that it will be possible to evolve SSBI into SSDS? This is debatable.
Data science requires creativity, curiosity, and an inherently technical and analytical mindset to be done well. Many would, therefore, argue that to produce projection data in the way that a data scientist does, only trained specialists will do.
However, with the growth of machine learning and its ever more sophisticated capabilities, running predictive analysis through an ML system and extracting easily-digestible, accurate data is becoming simpler. As such, there is a distinct possibility that machine learning tools will help to create a self-service data science culture that has the same benefit of allowing data scientists to concentrate on deep level work and enable business users to perform some data science tasks.
With a shortage of data scientists in the market at present, there’s every reason to expect that systems that hand over some work to business users to become more commonplace and as in-demand as SSBI. Nonetheless, like SSBI, a good business case needs to be established before picking a machine to do a human expert’s job.
If your business is considering a transition to or development of self-service business intelligence and would like advice on how to do this, do not hesitate to contact me, [email protected] I am also interested to hear your experiences if you have set up self-service business intelligence in the past, what challenges did you face? Do you have a great success story?