Advertisement: Bradfield Centre mid
Advertisement: TTP
Advertisement: RSM
ARM Innovation Hub
Barr Ellison Solicitors – commercial property
Advertisement: Mogrify
RealVNC mid banner careers
Mid banner advertisement: BDO
RealVNC mid-banner general
Advertisement Cambridge China Centre
Advertisement: Wild Knight Vodka
Advertisement EY mid banner
Advertisement: Cambridge Network
6 September, 2017 - 14:38 By News Desk

Companies urged to scale data capabilities to remain relevant

Geospock Steve Marsh

Dr Steve Marsh, founder of Cambridge startup GeoSpock, is an expert in solving big data problems at pace. He and his team are developing technology to solve problems for a world of autonomous vehicles and smart cities. Ben Sillitoe caught up with him to talk extreme data and building the tech of the future.

Big data startup GeoSpock has set itself the target of becoming the ‘Google for the physical world’ as industries, communities and people become increasingly connected.

It’s an ambitious business model that is effectively preparing for a future that is rapidly approaching: A future of autonomous vehicles and smart cities utilising real-time data from multiple connected points to improve the way communities and public services function.

GeoSpock founder and CEO Dr Steve Marsh believes it’s a world worth striving for.

The so-called Internet of Things (IoT) drive has the potential to reduce pollution and congestion on the roads and make towns, cities and countries far more efficient and safer places, he says.

The IoT space will comprise a series of device manufactures, he argues, who plug their systems into the physical infrastructure. Organisations will develop multiple applications to optimise bus timetables or manage traffic, for example, based on real-time data analysis of connected devices and people’s location.

It will create trillions of data lines per day and Marsh claims the management of that volume will only be achieved using GeoSpock’s technology.

“In this smart city app-marketplace world, a data management layer will exist which has the ability to harvest all data together and provide machine learning that makes sense out of the information for all parties – that’s where we’re angling to be,” he explains.

“Google gives you sub-second searches from the virtual world, but with our technology we can do sub-second searches of the physical world. Searching by time and location aligns very well for IoT use cases.”

Measuring this smart city ecosystem might entail data analysis of micro-climates, congestion or pollution and when combined with people’s smartphone data or vehicle GPS it builds a complete picture of the physical environment.

At this point and with the right tools in place, real-time anomaly detection can be utilised to fix problems like traffic congestion or pollution before they actually manifest themselves.

We don’t yet live in a world of smart cities and autonomous vehicles, but all the indications are that it is the direction of travel.

IDC forecasts worldwide spending on IoT to grow 16.7 per cent year on year in 2017 to reach $800 billion (£630bn). The analyst group expects global IoT spending to total nearly $1.4 trillion by 2021 as organisations continue to invest in the hardware, software, services and connectivity that enable an IoT world.

Marsh welcomes the fact the discussion has shifted away from the number of devices connected to the true value of software and services combining to enable the capture, interpretation and action on data produced by IoT endpoints.

He notes: “GeoSpock has been solving these problems for four and a half years, before IoT was a thing and autonomous cars were still in people’s dreams – but I could see this evolution towards an explosion of realtime big data being produced by sensors in the physical world.

“None of the other technology out there was built to handle that volume, velocity and variety of data. We needed to fix that fundamental problem because until then all of these future use cases, such as dynamic fleet control of autonomous cars and management of smart cities, will remain unattainable.”

Forget big data; how do you manage extreme data? IDC suggests worldwide revenues for big data and business analytics will grow from $130.1bn in 2016 to more than $203bn in 2020.

The rapid growth of this sector and rising demand for tools to address big data challenges is attributed to growth in information, a new generation of technology and a cultural shift towards data-driven decision making.

“Without using big data, you leave an awful lot of value on the table and the simple fact is in the future you won’t be competitive because someone like Amazon or Google will arrive and disrupt your whole industry,” Marsh says, adding: “It’s got to be a priority.”

GeoSpock’s technology goes beyond big data and moves into the realms of extreme data management and has the ability to shine a light on ‘dark data’ – unlike other search tools – according to Marsh.

Extreme data is that which is continually flowing into a system from generally non-human generated sources and the Cambridge company’s focus on this area represents its unique selling point.

GeoSpock enables “clever processing on the data”, Marsh explains, meaning relevant insights can be extrapolated regardless of how big the data pool gets. The company promises sub-second query performance no matter the size of information.

This strategy, the tech entrepreneur notes, usurps current methods of batch processing data, which can take hours, days and often weeks to compute. GeoSpock’s toolkit promises to turn data over in an instant, thanks to a geo-temporal referencing process.

“If your systems take ever longer to process data the more data you have, you’ll always be behind the curve and won’t be able to react quickly to new and changing situations,” Marsh argues.

GeoSpock’s target customers in the AdTech and telecoms industries are among those sitting on multiple Petabytes of data that would become a hugely valuable resource if better optimised.

For example, GeoSpock’s technology can help AdTech companies report back real-time, location-based digital ad engagement measurements, which enables advanced predictive and reactive campaign planning for their customers.
But, as is the case in multiple sectors including retail, banking and utilities, current datasets in GeoSpock’s target industries are only being used for commercial gain at a limited scale.

“We are now in a position where people have frantically hoarded as much data as they can get hold of without really knowing what to do with it at the other end,” says Marsh.

“You typically had to define up front the questions you wanted to answer because that was the limitation of the toolset. Organisations may have had a very different set of requirements when they started gathering the data and now there will be massive gaps in those insights.”

GeoSpock’s existing tools help users identify the data questions they need to be asking before answering them at an extreme scale. A machine learning component to the technology suite will help automatically find key trends in expansive data sets when it is launched next year.

Marsh is confident the tools are set to seriously change the big data industry and in the process establish a platform to deal with the future-connected world he envisages.

“GeoSpock has approached things from the extreme end first, knowing data is going to be generated at a much quicker rate at an ever larger scale,” he says.

“We’ve solved the problem beyond where most of the market currently is, but in two years everyone will be crying out and GeoSpock will be the only solution that can scale up to meet demand.”

The road to the future
GeoSpock has mapped out a clear route that it hopes will nurture its business towards its ultimate function in the predicted IoT-defined, smarter world.

That involves servicing the AdTech and telecommunications markets first before focusing its efforts on the automotive sector and smart cities when the time dictates. The order has been chosen for a reason.

“Mobile ads are generating up to 100 billion new data points every day for each company. Those involved in the industry are breaking existing systems and can’t get those insights out,” Marsh says.

“AdTech provides a great market and mechanism for us to go and prove ourselves. We are going to give them the ability to get sub-second insights on not just what they generate in a single day, but on what they generate across multiple years. This is because our system was purpose built to search trillions of data points rapidly.”

The CEO says this neatly blends into the second target market, telco, which is in the process of opening secondary revenue streams for advertising that require more sophisticated data management techniques. It is then his focus will turn to automotive.

“There are giant companies out there harvesting all this data but they don’t really know yet what the killer application is,” says Marsh.

“They’ll be looking for companies with extreme data expertise to partner with. We know that when we are able to prove to them that we can solve the problem at extreme scale in telco, then automotive won’t be too much of a leap for us in terms of size and velocity of data.

“It’s about credentialing yourselves – if you can do phones and car tracking then data applications for smart cities is the obvious next step.”

GeoSpock wants to reach the point where it is a facilitator between giant auto and telco companies, helping them create secondary revenue channels by selling their intelligence back to a smart city ecosystem. That is fundamentally what the company means by becoming the Google of the physical world.

“We started this company four and a half years ago to solve problems that didn’t yet exist,” notes Marsh.

“It’s all about understanding where things are going, and actually being able to better use the resources we have as a species is fundamental to our survival. We take for granted what we have, but bigger problems in the future will need resolving.”

• PHOTOGRAPH SHOWS: GeoSpock founder and CEO Dr Steve Marsh

Newsletter Subscription

Stay informed of the latest news and features