Imagine a programme that could tell you who in your CRM is most likely to buy.

You’re thinking you can already do that, but the question is how accurately can you do it? Not very, because the characteristics of likely buyers are a million miles from perfect.

I think it was 2013 when it was first said that 90% of the World’s data has been created in the last 2 years. Apparently, we create 2.5 quintillion bytes of data a day – I had to look up quintillion – it’s 10 to the 30; of course, that includes many different types of data such as sensor data, not just digital content or data stored in databases. That’s big data.

The idea that you can find correlations in large data sets allowing you to make predictions that can make a business more effective is very powerful. If I could marry up web data (LinkedIn for example) with my CRM data and find correlations that predict buying then I would have an extremely powerful tool. Many are trying to do it but there are many hurdles to be overcome before businesses, especially small businesses, can do anything with all that data.

Our digital heritage means that we work with clients in cutting edge areas of technology and as a result, we’ve gotten very good at selling new “inventions” to big businesses. There have been many examples of new ways to do things over the last 20 years: SEO, PPC, affiliate marketing, web design, eCommerce, web usability, UX, programmatic buying, mobile, location-based marketing, apps. Commerce and government have become much better at changing how they do things and adopting new practices and tech.

So what’s the difference with big data?

Not much, but its potential is much harder to realise. The things I mentioned above can be trialled, evaluated and switched on. Analysis of business data to find correlations with KPIs, and writing software that can use AI to predict and optimise that metric to save or make money, you can’t just trial that. For one thing, a human has to tell a machine how to analyse the data. To do that, a human first has to understand a data set which given the volume of data requires a lot of time. It’s costly unless you have a machine that can do it.

The other problem is that computers think in digital; bits and bytes, fields, dropdowns. Absolute values. Humans think in analogue; text, thought, opinion, nuance, subtext, tone of voice and so on. The World is analogue and the best nuggets of insights are in the analogue data, and to interpret that data is problematic.

What’s out there now?

We’re lucky to work with a software start-up that’s making huge waves by doing predictive analysis of people data. They’ve written intelligent software that can understand language, can model and correlate with business performance, and can learn from its successes. They’re making sense of all the data that companies have and it’s brilliant.

What excites me is the idea of applying this to sales. Could this mean that machines can find the best leads in your sales data? Most of the tools in the current market, such as Hubspot, Marketo and SharpSpring manage the interaction with the buyer and although they do surface hot leads, they do so based on the history of the interaction. Other tools such as reverse-lookup tools like Lead Forensics and Candy bring to the surface people who have looked at your website but it is still the buyer interaction that turns a lead hot. You don’t make sales by knowing who visited your site or if they have downloaded two of your case studies.

The future

What will happen when someone writes a tool that is able to analyse all of your database fields including language data combined with LinkedIn and other online data, and model correlations that indicate buying behaviour. It’s undoubtedly possible. Imagine a machine that could tell you who is most likely to buy.

Well, the good news is that it’s possible and you can do it. It doesn’t even cost much. But it’s a fledgeling field and you need someone who can use the tools that are out there because the machines are still learning.

We’ve been trialling a new tool and there have been some interesting results. We really do believe that there is tremendous potential in this area.

If you want to find out more about this then please do get in touch.