How Did We Increase More Than 50% In Sales With Only 5000 Customers?

datafying How did we increase more than 50% sales for an online store with only 5000 customers

Photo Credit = Matt Biddulph@Flickr

First, we hardly analyse data with less than 1 million rows. But, it is a good case study to demonstrate how we use data to make a difference. So, we got a junior analyst to jump into this project.

*Due to our non-disclosure agreement, we can only disclose a limited amount of information regarding the company.

The background

The client had an online store with 5000 customers. We had full access to 3 data sources including customer data, transactional data and email data.

The problem

The main problem was that some of the key metrics for emails were dropping more and more as they sent more emails. Mostly importantly, emails did not drive traffic and sales as anticipated. Why? They only sent bulk emails to everyone in their customer database. Right, sending the same message to all customers is always a bad idea. We know everyone is different. So, it was already quite obvious to us!

The approach

We ran a few QCs on the data to check if there were any abnormalities. It is always important to do some QCs before analysing the data. Then, we aggregated their transactional data to a single customer view. And, we identified their email behaviours like who opened what who clicked what and most importantly who purchased what.

The insights

We identified 3 segments in total.

Segment A

Sales primarily came from the top 30% of this segment. Most of them had spent relatively more, had more frequent transactions and had shopped more recently than the rest. Also, they spent the most on category X. This segment was the most engaged with their email campaigns. Most opens and clicks generated from this segment.

Segment B

20% of their base spent and visited somewhat the same as the most loyal segment above. But, all had become dormant for the past 3 months. And, more than half of them were disengaged with their email campaigns. Half of the segment did not open or click on their emails. They spent mostly on category Y.

Segment C

The rest of the 50% had only got less than 5 transactions and they rarely engaged with their email campaigns in the past 2 years. Obviously, these customers were more disengaged than the 2 segments above. They spent mostly on category Z.

The recommendations

Based on the insights extracted from the data, our junior analyst had recommended:

For Segment A – Design a targeted and personalised email offers to cross-sell Category X with other categories for sales uplifts as data suggest a direct relationship with email engagement.  Reward the loyal customers unexpectedly with a little gift.

For Segment B – Design a targeted and personalised email offers with a Call to Action to ensure customers in this segment to return within a specific time to drive engagement in emails. Cross sell Category Y with Category X in the hope to transform this segment to loyal customers.

For Segment C – Design a series of targeted and personalised win back campaigns with an irresistible killer offer for Category Z. Conduct an A/B testing on the subject line to measure which one do better. Keep track of the opens and clicks from the email data. Do not waste resources to email after 3 campaigns.

The results

Segment A spent 50% more as many purchased other categories as well.

Segment B spent 30% more and they visited 2 times more and 10% had become Segment A.

Segment C spent 10% more and visited 1.5 times more and 5% had become Segment B.

So, why do you still deliver the same message to all of your customers?

Related Posts

Leave a Comment

datafying How to squeeze more juice from your customer data for your online storedatafying How to do better email marketing