March 24, 2016

Actions! Not just insights.

What are real world examples of IOT Big Data Use Cases impacting Biz Outcomes?
Flutura keenly read the market signals from Houston, Tokyo + European markets and more specifically in the energy + engineering industries we are focused on.
After analyzing the data accrued over the last 24 months we have distilled it down to six practical IOT big data use case families.
We have heard a lot of marketing hype around IOT and Big data. At the end of the day it boils down to ONE POWERFUL UNANSWERED QUESTION.  
What are some of the measurable outcomes impacted by Industrial IOT use cases?  
This is the holy grail of IOT and Big data. While most IOT big data use cases focus on cost optimization, there are a couple of interesting use cases which we encountered which activated new revenue streams by tweaking the business model in the market place. For example, in the Oil and Gas
Industry Flutura encountered organizations creating new revenue streams by monitoring real time situational awareness of digitized wells as a paid value added service. Flutura helped a provider of Smart city solutions to monitor 11,000 government owned buildings in real time by watching their boiler, chiller, fire alarm and other ambient data in real time where the pricing model was per device.
In the deregulated REP markets, Flutura has encountered REPs offering new value added offerings like energy audits and asset refinancing for commercial industrial customers which are powered by smart meter based big data products.
Energy consumption is a very important lever which drives profitability in the industrial setup. Surgically targeting this business outcome can really make the business case for IOT based big data solutions as we now have the ability to mine granular state information and correlate it to energy outcomes
For example, we recently executed an engagement in Houston based fleet provider where we reduced average fuel consumption by 2% unlocking 65 million dollars in savings per year. We did this by digging deep into signals buried in operating sensor event streams and location related data which was then correlated to the fuel consumption data.
Another project we executed which was in the utility industry involved reducing average peak power consumed by residential customers by analyzing smart meter data across millions of households gathered at 15 minute intervals.  This pattern seems to be repetitive in multiple industry context where a single digit change in energy/fuel efficiency has a six figure impact on savings
Industrial industries are asset intensive. For every minute an asset is down millions of dollars are lost. Reducing down time of certain nodal assets is a use case which has a business case. Let’s take two examples we executed in this area to illustrate this theme.
We worked with a leading Spanish wind power generation company recently who gave us data regarding the performance characteristics of 2 turbines like main shaft rpm, grid voltage, slide ring temperature, reactive power, and rotor characteristics at second level. We unleashed Machine learning algorithms on terabytes of time series data a Hadoop cluster and successfully extracted signals emitted prior to a turbine breakdown A US Engineering giant gave us data from Electrical trip unit data (48 Samples/Cycle and 4 Cycles prior and post events along with power data – voltage, current , frequency ) which was then analyzed to spot significant differences in trip unit calibration which compromised asset integrity “Smelling” asset signals is again a recurring pattern encountered across industries.
In process and discrete manufacturing industries, it’s very important to keep defects below a certain threshold. A new set of possibilities is being enabled from granular data collected from digital factories. It is the ability to dig deep into second level sensor data to understand specifics of process states which increased defect density. Let’s quickly take an example of a project we executed for a US based electrical product manufacturer.
The product underwent a variety of operations and at each step of the operations sensors were monitoring viscosity, humidity, temperature data at second level and streaming it to SCADA based Historians. We were able to spot patterns using advanced machine learning techniques on a Spark based architecture which resulted in an 8 % reduction in defect density for the product
As the grid gets increasingly intertwined into all the assets, security becomes a very important consideration. Security forensics using granular event data can help investigators analyze sequence patterns exhibited prior to an adverse event happening. These digital signatures can be codified into a knowledge bank and watched in real time.
Pricing innovations using analytics is another area where we are seeing some powerful use cases blossom. For example, in the energy value chain trading happens in Energy exchanges and there is a lot of data generated from trades, competitive pricing information which needs to be mined to understand volatility patterns and to help time the market. Flutura’s energy data scientist are helping a major energy trade see those invisible pricing patterns to optimize millions of dollars by timing the market well by analyzing past correlations Another use case which is finding momentum is dynamic asset pricing. Assets which are leased to customers (for example golf carts are equipped with sensors which record location, topple events, average cart speed, start, stop events). By analyzing the granular usage statistics, asset leasing companies are able to have dynamic pricing of assets.
We do agree that the marketing machines have been on steroids advocating IOT and Big Data as the panacea for everything.  Industrial mindsets focus on tangible outcomes and some of the frequent questions we hear are they just fads or is it real?  Where is the business case for IOT big data in Engineering industries? What are the use cases which are real for the Oil and Gas industry?  How can Retail energy providers and energy trading firms benefit from IOT and big data?   While a lot has been done on buzz word introduction, we felt the best way to make the case was to highlight tangible business outcomes which are tangible and "show them the money".  We hope you found the use case taxonomy useful and would love to hear your experiences from the trenches. As we say in Flutura…   May the big data renaissance awake every industrial organization.


Flutura is a niche Big Data analytics solutions company based out of Palo Alto (Development Centre in Bangalore) with a vision to transform operational outcomes by monetizing machine data. It does so by triangulating economic impactful signals from fragmented data pools. The name Flutura stands for butterfly; inspired by nature's greatest transformation from a caterpillar to a butterfly. We are obsessed with Trust and Transformation and align our daily lives to these core principles. Flutura has recently been identified as one of the top 20 most promising Big Data companies across the globe by leading analyst magazine CIO Review. Flutura has also been featured on Gigaom reports on Big data + M2M in energy sector. Flutura was also the winner at Tech Sparks, where 800 innovative startups were evaluated.
Flutura is funded by Silicon Valley's leading VC fund The Hive (based in Palo Alto) which primarily invests in big data companies worldwide.

1 comment: