A KISS to deliver business value with Analytics

Some thoughts and observations around a K.I.S.S. (Keep It Simple Strategy) to deliver business value with analytics.  The examples used relate to my own experience but I believe a KISS (other acronyms are available) is applicable across multi-discipline organisations.

In terms of tools, every hardware and software vendor now has their own flavour of analytics (previously called reports/charts/graphs by the marketing folks) sitting on top of their silos of data.  There is also no shortage of analytic tool vendors, desktop, in-house server based, or cloud based. Exciting and exotic new names appear every day in the analytics space, promising unparalleled functionality.  Everything now supports a growing and bewildering set of BIG DATA technologies (Hadoop, Spark, Machine Learning, AI, R, Python ……) loved by in-house IT, the data science programming community and consultancies for big data development projects and requiring skill sets we are told we ‘must have’.

In Oil & Gas, upstream and downstream, there has never been any shortage of data or BIG DATA for that matter.  For more than 30 years IT has collected a vast array of application data lakes for engineers of all types: geologists, reservoir, drilling, petroleum, operations and maintenance, infrastructure projects and facilities construction, plus the strategic planning, forecasting, accounting and other mainstream business functions.  The potential of the digital oilfield has only added to the ocean of data.  What there has been a shortage of, is the ability to quickly deliver real business value from these data sources to all levels of this diverse community of citizen analysts.

As you build out your analytics strategy I would recommend you consider the following:

1.  Appoint an Analytics Leader

Ideally, an individual that can match business value with analytic potential and lead/advise/coach/mentor a team of citizen analysts/data scientists in attaining high value results.

2.  Identify Business Value versus Cost

During every oil price slump the leadership talk is always about making the industry more efficient but the actions taken are mainly short term related cost efficiencies to deliver shareholder value; delaying CAPEX, cutting headcount, reducing OPEX, and pressuring subcontractors to minimal margin levels and then wait for the oil price to rise.

This time round we have also witnessed the acceleration of ‘the great crew change’ and a cut in graduate recruitment which is hardly strategic in positioning the industry with the knowledge, experience and skills needed for the future.

Instead of the short-term market focus on the $$ value per BOE (Barrel of Oil Equivalent), there is huge opportunity to use analytics to help make the industry more efficient by a focussing on BOE per $$ invested.  There is no shortage of opportunities in oil & gas for continuous improvement and eliminating defects to deliver business value.  The KISS is to use analytics to turn insights into action, change behaviours and identify the actions that deliver the ‘biggest bang per buck’ invested.

With each KISS extreme business value can be identified and delivered through analytics on existing data sources without the need for a ‘BIG DATA’ technology project.

3. Ask Questions – Get Answers Fast

If you have already started down the traditional IT Business Intelligence model of tell me what you need and I will build your data warehouse, dashboards/reports/KPIs for you then you have probably spent considerable time and development budget on addressing the first two items (facts & questions) above.

Just like the gold miners of the 1890’s with their method to find gold, CLASSIC BI is a too-late architecture because by the time the dashboards / reports are delivered the business need has moved on.  IT development / support has little chance of keeping pace with answering the next question.

Fortunately, E&P IT has been capturing everything for decades in all these data silos and has the skills to deliver a platform to store, refine & analyse all data sources.  The KISS is to provide a self-service enterprise level discovery tool on top of this data to the citizen analysts on the desktop, browser, mobile and internet.

4. Meta Information Models and Data Marts/Enterprise Data Warehouse (EDW)

The advantage of the CLASSIC BI tools is that they usually have a meta information layer where the data is presented to the end user using real business names/terms and any calculations are validated through a testing process. The KISS for IT is to let the database do the work of delivering data using data table joins/views which have real business names and not the default tablename.columnname that the citizen analyst struggles to interpret.

The database and EDW vendors model of delivering master data management, data validation and processing, calculation/aggregation/forecasting/modelling at the database level is the optimal approach and the data appliance vendors have raised the bar on delivering data for analytics.  The KISS is that despite all the hype you do not immediately need an all-encompassing EDW to deliver data to your citizen analysts.

If you are building an EDW, a self-service discovery/insights platform will help define the ‘most popular/needed’ data structures needed to be incorporated into the EDW design to quickly deliver value back into the business.  For the self-service discovery citizen analysts, the EDW is an additional high performance, secure and trusted data source available to them.

5.  Data at Rest plus Data in Motion

Too many leadership meetings are focused on historical data supported by polished PowerPoint slideshows describing where we have been rather than where we are heading and the options available.

Blending historic data with near real-time data to publish/deliver ‘live’ analytics to drive meetings and daily decision making, helps management to focus on issues, past and future events, prompt discussion and develop options.

The KISS of using analytics and business data in this very visible way addresses the third and fourth items (intuition & exploration) above. This is a behaviour game changer in terms of moving decision makers from a reactive to a proactive culture, accelerating the decision-making process plus continuous improvement of data quality and value.

6.  Predictive Analytics and Event Processing

Field Monitoring – Which one is you?

In Oil & Gas upstream & downstream there have been many successful projects promoted by the industry on the right side of the chart above.  Sadly, in terms of global adoption the oil & gas industry is risk averse and ‘everyone’ is still watching big screen TV, believing this is the best available.  The industry can improve in the IoT/Smart/Mobile environment, but only if they are prepared to learn from predictive and event driven technology in other industries; manufacturing, logistics, retail and most of all aviation.

The technology exists to deliver a ‘future perfect’ asset which has minimal unplanned downtime and where equipment reliability and availability is predictable.  With in-memory DataMart’s of live sensor streaming data looking for time based events, for example, if you see A, followed by B …. then the probability of a failure C is high and you can automate some alert/action to prevent the failure.

There are already many successful use cases for this approach; corrosion monitoring, sensor re-calibration, rotating equipment start and stops/runtime hours, electric submersible pumps (ESPs).

One of the most interesting use-cases for predictive and event-driven analytics is to look beyond individual equipment items to the system and process envelope and to predict when a system is moving out of the optimal/safe operating conditions and take action.

The KISS here is that with minimal training with the tools, your own engineers with the knowledge and experience of the assets, and not some huge IT development project, can be building, testing and deploying these predictive and event-driven models in a matter of a few days per use-case.  Predictive models can be tested against historical events and continuously improved/fine-tuned to ignore ‘false positives’.

Predicting, taking action and thus preventing a production shutdown/outage is where analytics can deliver huge value.

TIBCO Spotfire has proven itself over a significant customer and user base in Oil & Gas to be an excellent enterprise analytics platform providing data governance and security.  Out-of-the-box (OOTB) TIBCO Spotfire provides a comprehensive analysis tool for the citizen analyst and data scientist alike.  If you need simple, Spotfire is quick and easy to use, able to connect directly and merge/blend with all oil & gas data sources.  If you have more complex data statistical requirements, OOTB Spotfire allows the advanced user to quickly build new analytic tools and workflows to deliver ‘Analytics for Engineers’.

Even with all the existing Spotfire use-cases there remains huge potential to build new and innovative analytic tools, predictive and event-driven workflows (Automation Services, StreamBase and Event Analytics), and insights to deliver extreme value for oil & gas from existing data.