Marketers get billions of data points from their customers, be it about which products were searched the most, whether they were added to the cart, did the customer read a review or checked if the brand delivers to a customer’s pincode, each and every action generates unique data points.
And these data points have never been this big in quantity.
Using this data, companies today are personalizing their communication with the
customers and creating hyper-scale personalized profiles for every customer.
To understand their customers well and use the data points
and insights to provide a seamless shopping experience to its customer,
Wakefit.co, a popular sleep and home solutions brand started looking deep into
their data.
When the scale of a business is small, it’s easy and
possible to make every decision manually. But when a business scales and has to
process, say, 5000 orders a day, it becomes impossible to take every decision
manually, while keeping the various business parameters in mind. As companies
scale, they need a higher degree of automation. For instance, Wakefit.co has
deployed systems and technology which mimics certain attributes of a marketer
and helps the marketing team make informed decisions.
“We look at customer engagement in two parts. One part
pertains to the acquisition of a new customer and the second part involves
driving repeat customers. We have serviced more than 800000 customers so far,
which is constantly growing, and by getting repeat users from this customer
base month on month or even quarter on quarter, we are able to see a lot of
incremental value in our business growth. To facilitate this, we look at
segmenting these customers and strategically targeting them to engage with the
brand,” said Puneet Tripathi, Head- Data Science, Wakefit.co.
With the data points that Wakefit.co collects, it is able to
recommend a better product to customers according to their search history.
“Looking at historical data, when a customer buys a product from
the website, we question ourselves if there was scope for us to offer that
customer a better campaign while they were looking for products. At that point
in time, through our pop-ups, was it possible for us to put a better product
proposition for these customers? We have worked on getting the answers to these
questions and have been able to offer a better experience to people visiting
our website over time,” he added.
Wakefit,co is now working on an intent model where the
company is trying to understand, with the help of past data, the intent of
customers, and the probability of conversion of these customers. This model
allows the company to proactively predict the customer sets who have a higher
probability to end up buying, because they mimic the behaviour of a customer
who generally converts.
Tripathi believes that this model is a gamechanger for them
because it doesn’t just give the company more insights on the probability of
conversion but also tells them what are the levers or features which leads
customers to conversion.
While the model has already been built for this project,
Tripathi said that it is still a work in progress because in the coming months
the predictive power might go down unless accurate data and insights are baked
into the model. Over a period of time, every model ends up losing its power
unless you maintain the model, retune it, and add more data and additional
patterns to it.
“Our marketing team has an understanding of how
the ROI of an upcoming campaign will look like based on the previous campaigns.
We are yet to automate this. But as of now, we keep helping them with the
insights that we have. We keep on hypothesising on a lot of new things to work
on.” Tripathi concluded. Courtesy: www.cio.economictimes.indiatimes.com
