The consumer internet creates a lot of data. Posts, tweets, photos, likes and searches. With over 100m photos and videos exchanged daily on Instagram, you get the picture! Big data connects data originating from one individual for monetary gain. 75% accuracy is normal but working at such scale ‘mostly right’ even ‘usable’ is good enough. However, imagine a B2B activity where 1 in 4 answers is wrong. More than a typo, that’s bad data. So in B2B data quality is just as important as volume.
Good quality data is commonly held as being accurate, complete, legitimate, relevant and current. When it comes to AI, add also granular (ie sufficiently specific to differentiate two situations), dispersed (to cover enough situations) and well labelled ie situation A correctly maps to output B. That seems a high bar until you know 100 images can be enough to train a visual inspection algorithm to check for production faults.
The conversation around digital transformation and manufacturing 4.0 informs part of the solution. More digitisation means more data. So the connected factory/IoT devices helps significantly in the production environment. AI can be trained to do what it does best – unpick a seeming jumble of data to find the patterns needed for decision making.
But this leaves one big area to address – what people do. This is very important in sales. Why did customer A buy but prospect B went elsewhere?
The trouble with people is that they don’t make good data sources. How consistent is behaviour? Who has the awareness and memory for perfect recollection? And that’s before we get to the willingness and time to record events as they happen and as they happened! So for most businesses the CRM system is solving the wrong problem – it is a place to put data, not a way of collecting good data.
There are three key repetitive activities at the front end of the sales process: needs identification, solution choice and price qualification. Repetitive activities make for good automation which in turn enables data capture. So one option is to take this part of the existing sales process and automate it. Digitisation of these information tasks yields other benefits: sales can be carried out digitally and so websites can both market and sell; human based resources can be focussed on qualified prospects undertaking subjective tasks such as relationship building and creativity. As with other areas of machine based automation, data collection can be built in.
With the data, AI can be trained and sales can be automated… but without automation there is no data – the data “chicken and egg”. This can be overcome by seeding the system with the elements under the manufacturer’s control – what can be made and which product best meets what need. Pragmatism is needed, as is an understanding that data is a journey. The results transformational. Imagine the benefits.
What could your business do with the answers to why customer A bought. Better still, what to offer prospect B so they buy from you too?