Prodsight

Reduce churn by analyzing Intercom and Zendesk conversations

Reducing customer churn is difficult simply because it’s hard to capture the reasons why customers churn. To get insight, you may opt to send out a survey to your churned users. However, when users churn they typically become disengaged are unwilling to invest time in providing feedback and usually provide a short, shallow reason as to why they churned or chose not to respond at all. This leaves you with little to no insight into why users are churning and it can quickly make you feel like reducing churn is impossible.

Have Intercom or Zendesk? You’re in luck!

If your company has Intercom or Zendesk you are potentially sitting on a goldmine of insight as your churned users have likely already made comments on their experience in your Intercom or Zendesk conversations. These conversations will contain issues raised by all of your users and not, specifically, churned users. Therefore, you need a way to segment and examine only the conversations from churned users.

To segment your churned users, you can set up a filter to show all users that had a Net Promoter Score (NPS) of less than six. Alternatively, you may want to filter on a custom attribute that is set up already and distinguishes between churned and all other users.

Once you have a subset of churned user conversations, the challenge is now to identify their experience issues. To do this, you could manually tag each conversation to categorise them into related topic groups.

However, if the conversation volume is significant, this will demand a lot of time. Not only that, the value of these insights is dependent on consistent tagging across your team. As more time is invested, the likelihood that the tagging will remain consistent is slim.

Therefore, it’s likely that you will feel that the time and resources devoted to manual tagging would be better invested in proactively reducing churn. If you can automatically understand what your churn issues are, it makes it far easier to implement solutions to those issues. This is precisely what Prodsight can help you do.

Segment churned users

We have built Prodsight to help you identify churn issues as efficiently as possible.

The first step in doing so would be to segment the conversations you have had with churned users.

This is done simply by selecting the NPS filter and setting a value range as you can below. If you do have a custom attribute that already distinguishes between churned and other users, then you will be able to segment using this.

Identify high-level churn reasons

Prodsight can automatically analyse all of your Zendesk and Intercom conversations for topics and sentiment. Once your filter is selected, however, we will surface only the conversation topic results from users that match that criteria.

Prodsight will then rank these conversations by the number of mentions and the overall sentiment. In surfacing the high-level issues, you can start to understand what problem areas are contributing to churn.

Narrow down specific churn reasons

After you understand the overall issues, you can begin to examine each groups’ corresponding sub-topics. Sub-topics provide more context and insight into the underlying churn issue within each group.

The sub-topic breakdown can reveal an issue in a core area of your product or service where users are set to derive the most value. This issue can then become a leading indicator of why this segment has churned.

To ensure that the results are as accurate as possible, Prodsight reruns the analysis every twenty-four hours.

Verify the impact

Once you select a high-ranking sub-topic, you can visually track the number of mentions over time and its overall sentiment. You can verify the effect of this sub-topic by correlating the graph with your monthly churn rate.

Let’s say your churn was higher in October and November and that the number of mentions of a  sub-topic was also higher. It is, therefore, possible that as more users experienced this issue, more of them decided to churn.

Using the graph can be a quick and effective way of discovering leading indicators of churn.

Understand why churn is happening

Now that you have verified the impact of the issue, you want to start addressing it. The first step in doing so would be to understand why the problem is occurring.

Understanding the problem is made easy with Prodsight as we surface all the conversation snippets that relate to the sub-topic you’ve selected. To help you analyse these snippets, you can combine phrases that are commonly mentioned together or enter a phrase you feel is relevant.

As you start to analyse these snippets, themes will become apparent as to what the root cause of the issue is. Once the reason is encapsulated, you can begin sharing the insights with your team and brainstorm as to how you might resolve the issue.


If you want to see how Prodsight can help you reduce churn, check out our free seven-day trial.