Additional Author #1: Gabor Mudra
Additional Author #2: Gergely Szerovay
Ordering: 1

1. Investing in algorithms has become a no brainer for large companies

Neural networks, machine learning or deep learning have now assumed their role as a center piece of corporate jargon. After IBM, Google, Facebook or Amazon, large industrial companies do their best to attract the best talent or acquire startups that have built strong algorithms

2. Garbage In, Garbage Out

Meaning that flawed, or nonsense input data produces nonsense output or "garbage". Algorithms are not magic black boxes from which compelling insights arise. Verified, structured data is the right way to start. In many cases, data has to be verified by human experts who are assisted by powerful analytic tools. From there, the best algorithms can be optimized.

3. Data is alive


Data freshness is only as good as when it was last verified, and this can be a few seconds ago, a few days ago, or a few months ago. In other words, the data about the data (i.e. “Metadata”) makes a difference.


4. Well maintaining, enriching and matching data is what makes it valued


Acquiring data is the comparatively easy part (even though it really is not so simple). Feeding massive quantity of data over time into your system so that data is actionable is the ultimate asset. This process has created strong market advantages for Google and Facebook, both of which have come to dominate the highly competitive global advertising market in less than a decade, dwarfing all other historical advertising players.


5. Passion at the top


Successful companies have hired early on a smart, passionate vice president of data and analytics and have their respective CEOs leading the effort (not the management but the CEO). Look at Jeff Bezos or Tim Cook as examples of this CEO-driven leadership. In other words, successful implementation stems from perseverance, rigueur and passion at a company’s very top.


6. Data is an asset


To be sure, this is common sense, and simple algorithms can bring plenty. But the gems are discovered by working the data and the algorithms over and over. Solving this jigsaw requires time, trial and error. Some initiatives will fail, others will net only small returns, and some unexpected ones will bring major return. Data’s return on investment is hard to evaluate upfront but experience shows that it is always worth the journey and ultimately pays off in aggregate.


We at AIEnergizer support energy companies with a successful introduction of Machine Learning into their business practices. Want to learn how? Contact Kasper Walet.


Source: Official Board