Videos          Articles

Big Data overload:
How to separate good data
from bad data?

The size of the datasets being created today is enormous, so new and sophisticated tools are necessary to analyse, measure and put this data to work. The world generates some 2.5 quintillion bytes of data every day, with IBM estimating that about 90% of the data available in the world today has been created over the past two years [1].

But experts warn that the overwhelming quantity of data, especially when it comes to company research, is not bulletproof and that a good dose of ‘realism’ is needed when sorting the wheat from the chaff.

This was one of the topics discussed when Citywire and BlackRock sat down together with big data connoisseurs to debate on the role big data plays in our lives and the process behind data gathering and analysis.

Nobody’s perfect

As an example of a ‘bad data’ event, panelists commented on the recent false judgment that a number of analysts made about the sale performance of Apple’s newest iPhone.

Since the start of 2018, reports citied several analysts and large banks playing down the demands of the iPhone – including the new iPhone X, on the back of their analysis of data coming from vendors of the famous smart phone. However, Apple’s successful second quarter results in May proved them wrong and brought the stock price back up after weeks of lows.

As many events in the stock market are difficult to predict, such as forecasting price movements, panelists said it is important to always put data sources into context, build confidence in the data and avoid rushed decisions.

Simon Weinberger, managing director and portfolio manager at BlackRock said:  “There’s a lot of noise in markets and you need an understanding of the market environment to put things into perspective and then make a call whether you want to sit tight, let it play for a bit longer or maybe something has changed structurally so the right action might be to actually change course and not use that particular source of data anymore.”

Brad Betts, managing director and data scientist at BlackRock added “We will argue around a particular data source, is this sensible? What might the indicators in there be? If patterns are being exposed algorithmically, do those make sense or not? We also do a lot of ancillary testing – if it predicts this, we think it should also predict that. We like things with no surprise and consistency.”

He said: “Our allegiance is not set on machine learning, artificial intelligence, natural language processing, our allegiance is to whatever is efficacious.

In support to this approach, Weinberger offered an example of a ‘good data’ source that was beneficial to the portfolio. He said the team had an overweight position in a German company on the back of the ratings of its employees across different factors such as management skills and strategy. “I really feel this is a great inside view into the company,” said Weinberger.

He said: “In this instance, this German financial firm are very much under the radar because it’s not in the main indices, therefore, as a passive investor you would have no exposure, full stop.  We did pick up on the fact that actually, employees were very bullish on the strategy.  They were very complementary of management and as such, that was driving an overweight position. 

“Now more recently, that company did present some very strong results and as such, we could convert that insight into our fund for our clients.”

Due diligence

As the quantity of data grows exponentially, some believe it is inevitable to apply an extra layer of due diligence when selecting funds that use big data in their processes. However, Anastasia Diangelaki, director at RBC Wealth Management said the premise of the due diligence on fund managers should be the same as for other types of funds.

She said: “You’re looking for strong teams, strong backing, and strong process that can give you some degree of certainty that the results that you have seen being produced can be replicated in the future.”

Obsessed with getting data’

When gathering data, experts say it is important to constantly question the approach.

Betts said: “We’re looking afterwards; what data did we not have that might have helped us in that process? Then we will go out and ask; can we find that data or find something similar?”

He added: “We are obsessed with getting data and getting as much as we can.”

BlackRock ‘consumes’ roughly three terabytes of data every day looking at about 10,000 companies globally. Through the various stages of the data analysis, Weinberger said the team runs robust field tests to detect any possible mistakes. For example, this might include a data provider that changed the formatting of the data which would damage forecasts.

“This is why you need the constant feedback loop between human and machines, to make sure that the process is robust and gets continuously improved,” remarked Weinberger.

[1] IBM Marketing Cloud

Big data and mixed messages: An example 

‘We have bunch of indicators; the more of them that are going green, the more you might long something.  The more of them that are going red, the more you might short something, and you have roughly equal amounts then you would have no active bet


‘We had a lot of lights that were indicating red and one light that was going quite green and the initial suspicion was that it had to be a bug.  It had to be and yet, you looked inside the data and you looked at what was being revealed and what was coming out of the algorithm and you realised there’s something going on that’s very different than what we thought. In fact, in that case, we bet with that indicator and it was a very profitable trade for us.’ 

Bradley J. Betts, managing director and data scientist at BlackRock

Capital at risk. The value of investments and the income from them can fall as well as rise and are not guaranteed. You may not get back the amount originally invested.
Past performance is not a reliable indicator of current or future results and should not be the sole factor of consideration when selecting a product or strategy.

Important Information
This material is for distribution to Professional Clients (as defined by the Financial Conduct Authority or MiFID Rules) and Qualified Investors only and should not be relied upon by any other persons.
Issued by BlackRock Investment Management (UK) Limited, authorised and regulated by the Financial Conduct Authority. Registered office: 12 Throgmorton Avenue, London, EC2N 2DL. Tel: 020 7743 3000. Registered in England No. 2020394. For your protection telephone calls are usually recorded. BlackRock is a trading name of BlackRock Investment Management (UK) Limited.
Changes in the rates of exchange between currencies may cause the value of investments to diminish or increase. Fluctuation may be particularly marked in the case of a higher volatility fund and the value of an investment may fall suddenly and substantially. Levels and basis of taxation may change from time to time.
Any research in this document has been procured and may have been acted on by BlackRock for its own purpose. The results of such research are being made available only incidentally. The views expressed do not constitute investment or any other advice and are subject to change. They do not necessarily reflect the views of any company in the BlackRock Group or any part thereof and no assurances are made as to their accuracy.
This document is for information purposes only and does not constitute an offer or invitation to anyone to invest in any BlackRock funds and has not been prepared in connection with any such offer.
© 2018 BlackRock, Inc. All Rights reserved. BLACKROCK, BLACKROCK SOLUTIONS, iSHARES, BUILD ON BLACKROCK and SO WHAT DO I DO WITH MY MONEY are registered and unregistered trademarks of BlackRock, Inc. or its subsidiaries in the United States and elsewhere. All other trademarks are those of their respective owners.