Gabriel Hughes is CEO of Metageni, a company that creates analytics for data driven decision making. Gabriel is an ex-Googler with a background in research, analytics and product development, and believes algorithms have the power to unlock the potential in customer data.

How important is data to the future of your organisation?

Data is potential. Data has within it the potential to unlock huge amounts of value for any organisation that is willing to use it to drive smarter decision making. I think being more data driven in the way that you make decisions is really the key to unlocking the potential that is implicit in data


How can organisations help to foster a more data- driven culture?

I think the answer to that question is about the whole ecosystem within an organisation. So yes, you need to get the engineering right, yes, you need to get the data integrated. Yes, you need to bring on data scientists, but really, the thing I think that sometimes gets missed out is, what is the willingness of the organisation to really adapt to become more data driven in its culture, and I think the cultural aspects of this is really important. It’s still the case in many organisations that decisions are made by the highest paid person or the person with the opinion- rather than letting the data decide. So I think you need a real willingness to listen to what the data is telling you, to work with the complexities and the shortcomings even, of the data in order to be able to make progress.

What advice would you give to anyone wanting to use data to transform their business?

I think it’s a mistake to try and push all of your data into a big data lake and try to link everything to everything else all in one big go- it’s much better to pick a specific challenge and go after that with data that’s reasonably reliable and try to launch and iterate and work in smaller increments towards your goal, because there will always be imperfect data, gaps in the data, data that isn’t integrated- it’s a process. It’s not something you can fix in one go.


What are the biggest challenges when it comes to turning data into actionable knowledge?

The biggest challenges when it comes to turning data into actionable knowledge are very often about trying to bring together the needs & priorities of decision makers on the one hand with the capabilities and skills of your data scientists and your data analytics experts  on the other. One part of the organisation has a top-down strategic way of looking at the problem and needs answers quickly, and the other tends to look at the data from the bottom-up, is cautious about making big inferences and big statements, is looking for statistical robustness and has a very high standard for what defines truth. Work on bridging that gap, and I think you’re onto something.

Data has within it the potential to unlock huge amounts of value for any organisation that is willing to use it to drive smarter decision making.

Gabriel Hughes

Should everyone be using Machine Learning & AI?

I mean at some level, ML & AI is just about pattern recognition. It’s just a technology which you’re leveraging through the systems and tools that you use. I think that companies should not be put off by the potential size of the investment required. So if you’re selective and careful about what you do and you pay attention, you can make great strides forward, I think with relatively few resources. The potential is huge and anyone who’s working with data should really consider investing at this stage because before long, it will be something they have to do, and not something that’s an opportunity- and they’ll be playing catch-up, and you don’t want to be in that position.

What’s your favourite piece of tech?

The open source libraries that are available for machine learning, particularly the Python Scikit-learn libraries. But also available within Python, I’m particularly excited at the moment about by the potential of genetic algorithms. Genetic algorithms are a machine learning technology that have been around for some time, but they’re still not widely understood or used, and I think they’ve got huge potential for optimisation where you’ve got a lot of complex possible solutions to a problem, a genetic algorithm can be quite an efficient way of finding the answer using evolutionary principles, you’re able to essentially search for an answer. They’re just really cool and it’s really great fun to use a genetic algorithm to try and solve an optimisation problem and then see what happens when you leave it, you come back and then BINGO! It’s found a new predictive model that  that predicts at a much higher rate than you thought was possible, that kind of thing. I think that’s a really cool piece of technology.


What are your main objectives for 2017?

To build a stable and scalable platform. We do have an early version of a platform that we are working with a select number of clients with which is a business intelligence platform which is about optimising the customer journey and which is driven by machine learning. That’s our challenge for 2017 and probably beyond into 2018 and beyond that.