How to bridge the data divide

‘BI is tough – disconcertingly so.And that is where so many companies have gone wrong’

Image for How to bridge the data divide

 

Data is divisive. From project failure to escalating costs, management demands for better information are simply not being met. Business intelligence projects are high risk, high cost and take too long to deliver essential insight.

Business leaders know that access to fast, trusted and reliable data is increasingly key – not just to competitive differentiation, but simply to keep pace with the competition.

Yet more than two decades into the data revolution organisations are now painfully aware that upwards of 80% of business intelligence (BI) projects fail.  That is, not just fail to deliver benefit – but fail to work at all. And yet the need for better information is so strong, so compelling, that organisations continue to invest in a vain hope that, this time, it will be better.

So what has gone wrong? BI is a victim of its own self-promoted success, in many ways.  Reporting and analytics tools, especially the current crop of highly visual, intuitive tools, are deceptively simple. They appear to turn base data, any data, into compelling and easily understood business insight at the touch of a button. The truth is very different.

>See also: Big data doesn’t come of age: 5 growing pains facing businesses today

BI is tough – disconcertingly so. And that is where so many companies have gone wrong. Believing the myth, hype and sheer misinformation surrounding BI, too many companies have assumed it would be easy  – they have underinvested, especially in people, and made undeliverable promises to the business users. The result? BI projects that demand ever more funding to deliver ever less corporate value.

Cost complexity

The cost of the BI investment is also a huge point of contention because most companies simply do not understand the total cost of ownership of these types of data projects.  Providing BI costs so much more than a snazzy reporting tool.

By the time a company has factored in the database, data quality, ETL (extract, transform and load) software and reporting tools, the typical cost for even a small, 50-person solution is around £200,000 per year – and that doesn’t include salaries, any required training or hardware.

By failing to understand the complexity associated with data extraction, cleansing and storage up front, companies find it impossible to make a valid business case.

Instead, far too many businesses have invested in a pretty visualisation tool only to discover that it does perhaps 10% of the job and without extensive additional investment in data quality and ETL technology, the tool is useless. At which point the realisation hits that the cost of people, time, money and systems work to address the rest of the BI project is untenable and the project is shelved. Another failed investment.

The success of any data project, especially BI, is always going to be compromised without accurate costing up front. Organisations need to be realistic. £1,000 per month on an analytics tool that delivers graphical Key Performance Indicators (KPI) sounds great, but consider the full costs: how many people are running the project, at an employment cost of £50,000 each per year? How much has the training cost, the servers, the database and the ETL tools? Go for full disclosure on the total BI investment and most companies will discover the true cost is closer to £25,000 per month.  What ROI is that delivering to the business?

Old technology

Of course BI can deliver value – massive, business-changing value – but it demands a different approach. The essence of the problem is that the vast majority of BI projects are attempting to achieve the impossible by using technology that was never designed for the job.

RDBMS systems that are designed for operational purposes are pretty useless for analytics.  Attempts to shoehorn BI and analytics into RDBMS operational projects were, in retrospect, always designed to fail.

New technologies and systems designed specifically to handle analytics have been in the market for a while – they are mature and proven.  With the right approach they can address the cost, time and risk issues associated with traditional BI projects. However, organisations still need to recognise the challenges of data cleansing and storage – even with the right technology investment.

Does it really make sense to make the massive internal investment in hardware, tools and, critically, people required to achieve BI success? Data experts are highly sought after – attaining and retaining good people is becoming a massive concern.

Even if the business gets the right infrastructure in place, the entire BI solution can be rapidly derailed when key staff leave. The reality is that most IT departments simply will never have the resources needed to deliver fast, accurate business insight.

In an era of cloud computing and everything as a service, isn’t it time to consider a different model? How much more compelling for a senior manager to simply define the business need and entrust the problem to a third party that already has all the technology components and the skills required to turn the base data into business intelligence?

To define a clear business analytical requirement, hand over the data and then receive, within a finite timescale, a dashboard on a mobile with eight different KPIs or real-time updates on sales in-store that reflect weather patterns or traffic flows?

By doing this, organisations can remove the risks, provide cost transparency, and take the business-driven approach that is at the heart of BI success. 

In a multi-channel world with the continuous creation of new streams and sources of data, organisations need to be able to experiment with data, change horizons and innovate at a predictable cost and with an acceptable level of risk.

>See also: How to lose money and alienate customers with outdated data insights

That is simply not possible with the old technologies and methodologies. Any business attempting to find value from BI with these outdated approaches is always going to be playing catch up at best, or at worst throwing good money after bad.

Organisations clearly need to get information from their data. But there is now a choice: continue to take a high risk, high-cost approach to BI and hope, with fingers crossed, that their project will be one of the 20% that delivers or adopts a business-led, service-based approach that leverages innovative analytics technologies.

With this new approach based on a business model that both de-risks and reduces BI project costs, organisations can get the right information at the right price from their data and get measurable business value for the business.

 

Sourced from Laurence Armiger, Zizo

Comments (0)