Home > Press Releases

How to turn your contact center into an Early Warning System

How to turn your contact center into an Early Warning System

Contact centres originally existed to service customer enquiries. They were built as a cost centre. Customer services was seen as one of those things you just needed to do, and efficiencies were around measuring calls handled by agents and little attention to outcomes unless you considered RFT (“Right First Time”).

The paradigm changed when it was realised that happier customers spent more and told their friends. This generates a tension between customer outcomes and cost to serve. Many are still in the transition. This second paradigm is enhanced by digital transformation where self-service and chat enable an easier experience for customers as well as efficiencies for operators.

Contact centres as CX early warning systems

The third paradigm is the realisation that the contact centre is not a separate entity to the business, but an inter-connected window into the operations of the business where customers have been motivated enough, for positive or negative reasons (and unfortunately mostly the latter) to contact the business to resolve some part of their customer experience. It can be seen as an equivalent of the ‘cost of quality’ for a manufacturing business. Let’s call it COD or Cost of Deviation from customer journey.

When viewed in this way, there are several opportunities for recognising the COD. For example if we think about a telco which has various enquiries, some service recovery will cost the business more than others and some will have more of an impact on customer satisfaction. For example resetting a password might cost just the operator time of £5, but correcting an invoice might cost £50 and obviously something more material such as a handset issue or network issue will have a more material impact. By adding the true costs and weights, even if only approximate, gives a better picture of the business. This model can be further refined by comparing survey results with contact centre enquiries for different issues to see if there is a ‘multiplier’, for example for some issues such as mouldy fruit from a supermarket, some customers might just throw away the product and not call in, and of those, some might be more motivated to shop elsewhere whereas others may carry on if it’s not recurrent. From surveys, it might be calculated that the multiplier for mouldy fruit is 10:1 i.e. 10 people are affected for every one person that calls in. The ratio might change for a more expensive item or for a more serious issue such as food poisoning. So, after pulling this together, the cost of the contact centre might be £100m per annum, but the service recovery costs in terms of rectification, credits and churn might be an order of magnitude greater than that e.g. £1Bn pa. Fixing these different issues will present different challenges, particularly if they require cross-functional processes or the costs aren’t easy to identify in any one department’s P&L. However there the first rule of control is that you can’t measure what you can’t manage so having an Early Warning System (“EWS”) of the issues and their true COD will give you the head start to take decisions.

PrediCX is one system that acts as an EWS for a contact centre and can be customised to measure COD by identifying the fundamental topics from all of the unstructured as well as structured data. It can also be tuned to be able to forecast COD as well as measure current, as well as identify possible explanatory factors. So for example if some categories are small, but growing quickly then these can be prioritised, and there might be hints as to the root cause e.g. geography, cohort or demographic of affected users. Patterns can also be split out such as seasonality from issue ‘spikes’ from growing trends. For spikes, these can be identified within hours to feed into a ‘war room’ situation to add actionable and timely data to a high-pressure, timely resolution, as well as damage-limitation actions to the likely affected customer cohorts. For medium-term trends, these can be addressed in a strategic way based on RoI. In all situations, the cost can be accurately estimated both for impact and for rectification. In some situations this can also help with insurance, risk management and provision accounting. With the figures about it’s easy to see that a 10% saving of the cost centre is nothing compared to a 10% saving of the COD.


Viewed in the right way, the contact centre is much more than a cost centre, or even an asset to be optimised in isolation. It can provide an early warning system into an organisation, and be used to both qualify and quantify the otherwise hidden costs of customer journey failures.

Another interesting conclusion is that, ironically, by making it easier for customers to enquire or complain (even to an automated process or chatbot), you capture more actionable information.


Warwick Analytics

Warwick Analytics has developed PrediCX, the world’s most powerful machine learning platform for text and voice of customer data. Ideal for contact center automation and insight.

The company’s software is based on sophisticated computer algorithms developed at the University of Warwick over a decade of academic research.