Designing an auto-chasing system — A case study

A series of texts and emails that are triggered based on specific user actions

Damodar Badhwar
4 min readApr 7, 2020

--

About the business

‘BillTrim’ lowers your household bills (Internet, electricity) by negotiating on your behalf. That means customers pay less for the same plan.

snippet from the website

Who is this for

People who have visited our website but did not perform the intended action.

We call them “leads”.

This system is intended to trigger text & emails based on the specific activities of the leads.

The job of these messages is to educate the users about their problem (overpaying on their bill) and pitch in our solution.

The problem — A lot of volume, less man-power

It’s a good problem to have.

In our previous process, CSR (customer success rep) team used to chase the leads manually.

As we scaled, the need to automate this process became a necessity.

  1. Bandwidth — With 500+ registration a day. The team is getting short of hands.
  2. Human errors — They often forget to replace the variables in the templates.
  3. Measuring the effectiveness of email — No metrics to measure the performance of emails.
  4. Repetitive for CSRs — No time to even think innovatively. They’re stuck in a rut

Chasing process (Previously)

  1. CSR picks the leads from a kanban board
  2. They pick up the relevant message from their repository
  3. Selects relevant channel (email/text) and sends the message
  4. Repeat for all other leads

Enters the automated solution

The first thing was to identify all the stages where triggers can be placed. We identified these stages in the customer journey by:

  1. Analyzing the drops in the funnel
  2. Talking to the CSR team
  1. Abandon signup — Leads who left the onboarding process for various reasons.
  2. Served — Customers who got the service.
  3. Post-sales monitoring — Think of this stage as ‘under warranty’ period.

Setting up the campaigns & triggers

The basic architecture of a campaign is shown

Writing the sequences

While writing the sequences following things were taken into consideration

  1. Stage of awareness: Since our idea is very new, most of our users are not even problem-aware. Meaning they don’t even know that their bills can be lowered without changing their plan.
  2. Use the same language as our customers: We analyzed Facebook forums to come up with the topics that our customers (potential) are talking about.
  3. Resolve all the doubts in their head: We analyzed 350+ reviews and social media comments to come up with themes for our sequences.
  4. Copywriting techniques: We took a course on copywriting techniques to understand how we can use the gathered data in the best way.
Factors while writing an email

Benchmarks

  • Open rate: >20%
  • Click-through rate (CTR): >4%
  • Conversion: >2%
  • Unsubscribe rate: <0.2%

Early Results (28 days)

Quantitative report for overall system

Insights

  1. Conversion is better with text campaigns
  2. High (>2%) unsubscription rate after 3rd drip
  3. Morning drips are more effective
  4. Overall unsubscription rate is high

What’s next

Pretty happy with the reward vs the effort.

We handed over the process of testing new copies to the CSR team. As a product owner, it's quite easy and clear to tell what works and what doesn't.

The next improvement can be done to put more sophisticated triggers that can analyze the reply and send a relevant message.

Have questions in mind? Let’s connect & talk https://www.linkedin.com/in/meetdemo/

--

--