Reinvigorating the CRM practices operators use and modernising their approaches using Artificial Intelligence and Machine Learning to deliver increased retention is what we do best, but how we accomplish this is perhaps less well known.

Prior to discussing how we work, let us first look at what, until recently, has been considered the “norm”.

The Common Approach

Most marketing teams operate with a manual workflow which looks something like this:

1.  Define new offers
2.  Manually send them to segmented groups of customers'
3. Analyse the results on these groups
4. Compare this against existing offers
5. Pick the best performing offers to send from the analysis in steps 3 & 4

We have identified three flaws with this workflow.

The most important of these is that this workflow isn’t automated. Multiple teams may oversee the different steps which can lead to inefficiencies and escalating costs.

Secondly, the process of improving offers is usually a one-off piece of work which needs to be repeated by teams every few months.

Thirdly there is no statistical confidence in the analysis. You may see a type of offer perform well but how can you be sure this wasn’t down to outliers in the data? Statistical methods must be used to make informed decisions.

What Ibex does differently

Ibex approaches this in a fundamentally different way using machine learning models; our workflow is outlined in four steps.

1. Define new offers
2. Automatically test these offers on groups of customers'
3. Automatically analyse impact
4. Optimise

At Ibex, we have built deep learning networks which predict the likelihood of depositing for each combination of customers' and offers.

This application of machine learning, combined with over 20 other models, allows us to build a very accurate P&L statement for each customer and offer they could receive.

With the capability of automatically creating recommendations and sending them to customers, Ibex can choose the offer which has the greatest impact on a customer’s lifetime value.

Once offers are defined, Ibex enters an experimentation period. This is step 2 of the workflow where Ibex automatically tests which offers work best for groups of customers – no manual scheduling is required and this allows Ibex to collect data on how the offers perform for customers with different demographics and activity.

During this experimentation period, Ibex is simultaneously analysing what impact the offers are having (Step 3). This analysis incorporates statistical methodologies so we can assess the confidence in our models and predictions.

The final step is to optimise, where we continuously update our predictions over time as trends change (step 4).

Key Benefits

Deep Learning

Underpinning this whole workflow is our deep learning networks developed by our in-house data scientists. Multiple factors are taken into account such as the customer demographic, their deposit history, what device they use to access your platform, previous responses to marketing and more to deliver an offer that uses the correct channels of communication, at the right time using an offer that resonates with the customer.

This level of personalisation increases retention rates and is made possible by predicting the likelihood of depositing for each combination of customers' and offers.

Automation

This whole process is automated resulting in no customer segments needing to be defined and no manual scheduling.

Furthermore, the analysis work is already taken care of with this approach, reducing the workload for your team.

If your customer base significantly grows this would have traditionally caused a significant increase in costs as there is more execution work and more analysis required. Ibex can very easily scale as you acquire more customers, keeping your costs down as both the execution work and analysis is already taken care of.

Continuous Learning

One of Ibex.ai’s key aspects is that it is constantly learning, the AI does this by frequently experimenting with a small number of recommendations so that we can continuously learn with fresh data.

This encompasses new offers or types of content that may engage your player base more thus improving its overall knowledge. This evolution of Ibex’s knowledge base is distinct from other CRM practices and allows us to pinpoint each individual customer’s needs through the knowledge gained across multiple data points.

‍

‍

‍