Predictive analytics make your email marketing relevant

The price difference between sending out 100,000 emails and 200,000 emails is pretty small. Maybe that’s why we’re all receiving so much rubbish. Looking at my inbox today, I see that most of what I receive is useless:

• the content is mostly irrelevant
• the frequency of receiving certain information is too high (and the content may be irrelevant, again)

I could always opt out – a pain for the sender because then they can’t use email to reach me, but also for me because I may miss something I need to know.

Individual messaging

Still, email marketing is a great vehicle for addressing consumers. It works best when the database and tool for email campaigns are not the end, but only the beginning of the project. What’s essential is a detailed understanding of the target audience. Individual messaging is a must and it has become a reality thanks to what is known as predictive analytics.

Predictive analytics relies on continuously collecting as much information as possible about everyone in your database from sources such as internal company data, website analytics, external databases, government open data, internet and social media data and all past interactions. The more data sources, the better. And by the way, there are more data available outside your company than in your ERP and CRM systems.

First movers have proven the concept

Statistical modelling based on these massive amounts of data will help you find patterns and predict what is relevant to your customer. If you believe that statistical modelling can help to predict customer behaviour, please read on.

The large European banks, mobile operators and retailers have hired data scientists to help them to introduce new services faster, lower churn and optimise their supply chains.

Thanks to predictive analysis, a retailer can now predict about 80% of the products that will be in about 80% of their customers’ shopping carts the next time they buy from them. Fantastic for inventory optimisation, but also great for sending out tailored promotions.

Predict the next purchase

It is now possible to have a precise idea what message you have to send to a customer, and when to send it, to influence their next purchase. You can’t do this with optimal campaign management – campaign analytics is what you need.

On a final note, data scientists are neither clairvoyant Madame Soleils nor Big Brothers. They cannot make fail-safe predictions at the individual level. We’re talking statistics here, so you should expect statements like: “There is a 85% chance that customer x will switch to another mobile operator in the next few weeks. Let’s send this specific email and follow up their activity.