Why set up an automatic customer feedback semantic analysis solution at Orange?
Our Business Services department was experiencing hardship in identifying satisfaction levers and improvement areas through our survey’s close-ended questions.
The fact that a solution performs the analyses is important. Before starting our collaboration with ERDIL, we had carried out an exercise called “Human vs. ERDIL”: a manual categorisation of the verbatim by all collaborators. We thought at first that the solution was inaccurate as it didn’t classify the same way as we did. We then looked at each verbatim and realised that many among us had carried out their classification in different ways. Analysis with the help of a solution may not be perfect but it comes with the advantage of being consistent, a very important factor when it comes to monitoring the evolution of our KPIs.
This step has convinced us of the relevance of the automatic analysis; it has also allowed us to focus on actual customer expression.
What are the advantages brought by ERDIL’s automatic analysis?
With an automatic solution like the one provided by ERDIL, verbatim categorisation and identification are carried out in a homogeneous way. We are consequently able to take into account topics freely expressed by our customers. We are also able to assess their impact and prioritise which actions are to be planned. We can also monitor the evolution and the transformation of the expressed topics over time.
Lastly, the implementation of automatic verbatim analysis has allowed us to lighten our surveys by avoiding irrelevant close-ended questions or by removing scores associated with a too large range of topics.
Feedback analysis allows to “quantify what’s qualitative but also shed light upon what’s quantitative”. When we observe certain developments in curves or score differences, we look in Esatis if there are some verbatim which could illustrate those variations.
How to encourage customer expression?
At Orange, we used to focus on KPIs, figures, scores. The addition of open-ended questions quickly allowed us to realise that this focus should shift towards the Voice of the Customer: what’s important is to be found within verbatim.
We are convinced as well that we have to perform segmented analyses that take customer expression into account. When planning targeted actions, a mapping of the different “battlefields” covering, for instance, activity sectors or satisfaction level, is carried out.
Can you provide some concrete examples of insights, key figures or ROIs gained with the help of the Orange customer feedback analyses?
Our agency managers specialising in customer listening manage to convey key messages to their Executive Committee with the help of analyses performed through Esatis, in turn allowing for easier action implementation.
Cross-checking various data stemming from Esatis and our information systems (IS) allows us to identify breakpoints and excellence levers in our customer journeys.
Analyses also allow us to look beyond the mere satisfaction/dissatisfaction opposition. In our survey, we ask our customers what they have thought of our services (with the NPS) and about their potential recommendations, based on their experience with other companies. Depending on their answers, we can identify our brand’s strengths and weaknesses compared with competitors and vice versa. This is yet another possible exploitation of the analyses.
How to get verbatim and scores to work together?
We also make use of the NPS in our surveys, which is a good metric but is rather disconnected from our employees’ daily experience. Verbatim on the other hand display a symmetry with their own experience on the ground. We have found a lot of connecting points with verbatim, a lot more than with surveys. A grade won’t ever have this power.
It is also important to maintain the link between verbatim and their author; an order number can allow the identification of a customer before calling them back to offer them corrective measures.
What did customer listening teach you?
The way customers express themselves regarding a same issue differs depending on their satisfaction level. The most dissatisfied customers are often more talkative. Detractors would rather express themselves on customer service basics (such as ease of contact, complaint handling) while promoters would tend to highlight human relations. There is thus no symmetry between detractors and promoters, hence targeting actions which imply different action plans and levers.