As we
discussed in my previous blog, CX Lifecycle Management for continuous
improvement in Customer Experience and revenue/customer,
https://personalandprofessionalexcellence.blogspot.com/2019/07/cx-life-cycle-management-for-continuous.html?_sm_au_=iVVWwvDnr4MNHqqF, Customer
experience follows life cycle similar to product life cycle and we can extend
this lifecycle through right interventions at right time. We also discussed
that time for intervention is critical and determined by the time when Customer
Experience Fatigue sets in. Just to refresh, Customer Experience Fatigue
is defined as the time and stage when the experience, which once thrilled
customers does not excite them anymore and they start looking for something
more, something different. This provokes customers to try out a different
experience offered by competitors and company start losing customers along with
associated revenue. The fig. below
depicts the CX lifecycle and the stage where Customer Experience Fatigue Kicks
in.
Critical
questions that come to mind are,
·
Can we
detect this fatigue?
·
Can we predict
it much before it sets in?
·
Do we know
reasons for the fatigue which we can address in the next intervention?
The answer
is YES for all the three questions. I believe that the recent developments in AI
& Machine Learning can help us to detect this fatigue earlier than it sets
in. Let us look at two scenarios, first a physical intervention at the store
location and second, an online intervention through the website carried out by
the company to improve their customer experience and see how AI could be used
to predict CX fatigue before it sets in.
Instore
Intervention: Let us
assume that a Retail / Grocery store introduces a “Creche for pets” at its
stores so that the customers can leave their pets at this creche and do their
shopping without any worries or time limits. To sweeten this initiative, the
stores does not charge any money for this service for all their loyalty
customers for certain time (could be little more than the average time taken by
a customer to shop in the store so that he/she can do some small errands beyond
grocery shopping also) and charge fees beyond this time. Customers who do not
have loyalty card, either can create one on the spot or can pay a deeply
discounted fee per hour.
The idea
here is that the customers can spend time leisurely in the store and can browse
through additional isles/ products and thus buying something which he/she does
not buy normally increasing the revenue. This initiative will also attract new
customers to the store and generate additional revenue. The incentivization
through Loyalty cards (which is typically free) will help us collect/analyze
the data about customers, their buying patterns, choices, other habits like
typical basket of customer during every trip and many more.
Proof of
the pudding is in eating, and ultimate reason for introduction of intervention
is to increase revenue. So, if revenue behavior can be modelled, one can
predict the future behavior, thus give an early indication of fatigue. The
model could be built using input parameters like footfalls in pet creche,
existing v/s new customers, time spent in the shop by customers using pet
creche or otherwise, shopping basket of loyal customers using pet creche and
before pet creche was introduced, Shopping
preferences and time spent in different aisles,
age, gender, race, fitness, etc. obtained from loyalty and computer
vision through multiple cameras in the store. Overall revenue can be considered
as output. The system can create Neural Networks to build the relationship
between various input variables described above & Revenue and tweak it continuously
as it learns. It will help predicting the revenue trend. If this model is
complemented with Information from social media interactions, the model could
also factor in the impact of social media chatter on the experience and
ultimately on revenue.
Website
intervention: Let us
assume that a retailer wants to introduce an online only clothing range for
adults and young adults (male and female) and creates a feature on the website
using computer vision so that the customer can try out a dress virtually. With
click of button, they can see the image of them wearing the selected dress as
if they are trying out the dress and looking themselves in the mirror. Company
also wants to provide a special service at a charge to alter the dress to some
extent and create a super fit dress for the customer at much cheaper price than
a tailored one. Silver lining could be added through a discount for trying the dress
virtually before buying it through the site.
This will
eliminate the hesitation in the minds of customer of online buying without
trying it out as well as almost eliminate the hassle involved in
return/exchange. A BIG relief for customer and lots of savings for the company.
Expectation is that after a big splash in the media and social media regarding
this intervention, it is expected that existing customers will try this new
feature and buy some more cloths. It will also attract new customers who are
attracted to the convenience and features provided by this intervention and buy
new cloths. Revenue will show a positive movement.
Similar to
earlier example neural networks could be created using inputs like cloth purchases
after using the feature and without using the feature, customers purchasing
cloths using this feature v/s those otherwise, Customers extending their
purchases beyond cloths to other items available on the web site, age, race,
gender, region, number of trials before a buy, etc. and out put as revenue from
online cloth sales. This will help predict the revenue behavior in future to
identify the fatigue in advance.
Once the
company is able to predict the CX fatigue, they can get another intervention
ready for introduction at an appropriate time. Company actually can build
another algorithm using results of multiple interventions and their impacts vis
a vis customer characteristic to identify the features that the next intervention
should contain or if the intervention has certain features, how will it impact
the revenue.
In
summary, Customer experience has a life cycle of
Introduction-Growth-Maturity-Decline. Every intervention created by company
will flow through these phases and will stop giving any results in due course.
An injection of new intervention at right time can restart the life cycle and
enable companies to maintain higher customer experience as well as higher
revenue per customer. AI can be very useful in predicting the intervention
point early through tracking of CX fatigue as well as defining features for
next intervention. Thus, company can continuously extend the CX lifecycle and
maintain positive trend in revenue.
Ref for
reading:
Customer Experience
Improvement using Differentiated Needs Pyramid (Maslow’s theory)
Predictability
in Customer Experience Improvement – A Perspective for Grocery Retail Industry
