1. Proving return on investment (ROI) to management may take time, but is feasible
Companies using predictive analytics in unconventional wells have often been able to tie its use to demonstrable gains in improved drilling accuracy, efficiency or cost savings. Doing the same for predictive use in maintenance and safety is less straightforward when avoidance of incidents or downtime are the main benefits. Efficiencies from these can be quantified, however, and 2–3 more years of running and validating algorithms and accumulating databased evidence should help many firms to demonstrate returns they have difficulty proving today.
2. Predictive analytics is delivering benefits today
Based on publicly available evidence, 57 of the world’s 100 largest oil and gas firms are using, or have plans to use, predictive analytics. Over half of these companies (34) report that their use of such tools has had a positive impact of varying degrees. Within the top 100, evidence of predictive adoption and use is most extensive upstream, in oil-field equipment and services, exploration and production. It is the largest firms in terms of revenue and market capitalisation (mainly integrated oil and gas companies) that appear to have advanced furthest with the technology.
3. Fields of application are widening beyond reservoir performance
Several large oil and gas companies have put predictive analytics to good use upstream, earning returns from improved performance of unconventional wells. Maintenance, safety and production control are newer areas of adoption. Companies are exploring future uses of predictive tools in combination with emerging technology capabilities such as machine vision4 and behavioural analytics.
4. Use cases in other industries may hold potential in oil and gas
Industry firms are scanning the health, aerospace, financial services and other industries for examples of predictive analytics use in equipment diagnostics, prognostics and maintenance, pricing and risk management. Other relevant use cases include machine vision in the automotive, sport and healthcare industries; behavioural models in the retail and financial services industries; and enhanced route optimisation used by transport and logistics firms.
5. Data quality issues hamper predictive analytics use, but these can be surmounted
Data quality and availability are viewed as the chief impediments to the use of predictive analytics in the industry. Oil and gas companies collect masses of data, but often can only access a small portion of it. Standardising disparate sources of data can also be difficult. Experts say, however, that data quality need not be perfect, and that workarounds are possible. Using advanced data science, for example, companies can filter data to overcome quality problems.
6. New data sources are gradually becoming available
Accessing the external data and expertise needed to make predictive analytics work requires oil and gas firms to widen the scope of their interaction. For example, the world’s technology giants support open platforms for developing algorithms, as well as data lakes and cloud platforms that provide great analytics power. Some executives interviewed for this study say their firms are benefiting from access to data and expertise on such platforms.
Download the full report and find out more about the pace, development and adoption of predictive analytics in oil and gas.