How Big Data Drives Investment Decisions
How Big Data Drives Investment Decisions
Although investment has traditionally been regarded as an art as much as a science, it has always been driven by data. The skill has been in how investors have interpreted the available data to gain an edge. Today, big data and machine learning are taking this skill to a higher level. Investment houses are now making decisions about where to put their money using AI-powered apps and terabytes of data. As more and more data is produced and the technology advances even further, could machine learning end up replacing the human investor?
What Kind of Data Is Available?
While still having access to the kinds of data on investment opportunities that they have always had, such as annual reports and profit forecasts, the advent of big data means investors can find out so much more.
Examples of new data that investors can analyse include:
- Credit card usage data – revealing consumers’ spending patterns.
- Social media data – allowing investors to see customer opinions in real time.
- Satellite imagery – investors can see how remote business locations such as oil rigs and mines are working.
What Can Investors Do With This Data?
Sophisticated software can collate an almost infinite amount of data, which can then be analysed by machine learning apps. These programs can spot trends in the data, which can be used to produce a wide variety of outcomes. For example, you could use machine learning to build models that:
- Predict how specific events affect areas of investment – e.g. If interest rates were raised to 4%, how would it affect the FTSE 100?
- Identify what drives movements in markets – e.g. Is the US stock market on a high because of energy prices or the price of the Dollar?
- Work out how to respond to even the smallest market moves – e.g. If BP shares fall, could you make money by purchasing ExxonMobil?
How Can Investors Use This Information?
When their machine learning programs generate predictions as mentioned above, investors have a choice of how to use them. Investment houses are experimenting with different models, each depending on the amount of human input. These include:
- Advisory role – Machine learning performs all the analysis and makes recommendations, but the final decision on whether to execute lies with a human being.
- Total autonomy – machine learning has total control over investment decisions with no human input. It decides what to do and executes based on those decisions.
Some investment institutions are already using these methods to drive investment decisions.
Alphabet Inc runs a venture capital arm called GV (formerly Google Ventures). It uses machine learning in an advisory role. Known as ‘The Machine’, the machine learning program decides a company is worth investing in, based on large amounts of data. These data points include market information, funding rounds achieved and partnerships. It then makes a recommendation using a traffic light system: red is ‘Don’t invest’; amber is ‘Wait’; and ‘green’ is ‘Invest’. However, this system has not been running long enough to know whether it has been successful.
Another VC firm, Hone Capital, uses a similar system to make decisions about whether to invest in startups. Hone Capital says using machine learning in this way has dramatically increased their deal flow.
Of course, nothing is perfect. There are some who believe the new machine learning approach isn’t necessarily the right way to go.
The art, rather than the science, of investing sometimes comes down to individual investors making decisions based on gut instinct which turn out to be wildly successful. They read between the lines in the data. Machine learning doesn’t have a ‘gut instinct’ feature. Until it does, will investors miss out on potentially lucrative opportunities?
Another fear is that machine learning programs like ‘The Machine’ at GV could be gamed. If you know what these programs are looking for, you can present your investment opportunity in a way that will tip the balance in your favour.
By its very nature, machine learning performs better the more it is used. The more data you feed it, the better the outcomes. PwC predicts that in 2020, there will be 20 times more usable financial data than there is today. For this reason, more institutions will be using machine learning to make investment decisions in the future. Investors who start early on machine learning will gain an edge.
Here are some companies we’re keeping our eye on as potential disruptors in the AI-driven fintech arena:
Oaknorth – https://www.oaknorth.com/
OakNorth is a UK bank for small and medium sized companies that provides business and property loans. The bank lends between £500,000 to £40m to businesses and property developers. Its loans have supported the development of 8,500 homes and creation of 8,000 jobs in the UK.
Cleo AI – https://www.meetcleo.com/
Your AI pal that looks after your money. Budget, save and track your spending.
Behavox – https://www.behavox.com/
Behavox is an AI-driven platform that transforms behavior in the financial workplace.