Atomic logo

Best Practices for Re-engagement

Re-engaging customers who are unsuccessful in converting can significantly impact your bottom line.

You can and should design systems to thoughtfully reach out to customers who have been identified as having high intention to convert and suggest they try again. People get distracted. Things come up. A well-timed message to retry can be a welcome reminder.

This guide will walk you through:

The first step is to identify unsuccessful customers with high intention to convert and group them into categories. Present them with a solution to the issue they faced and they will likely be able to complete the flow.

We'll begin by grouping the types of high-intent to understand the various ways we can spot these customers in the data. This is known as segmenting. We will group them based on the data we used to discern their high intent.

  1. There were multiple unsuccessful attempts.
    • If the end user initiated Transact several times, but the end state is is not Viewed Task Completed Page; there is a problem. This indicates high intent for this individual to convert but something is causing them to be unsuccessful. Gaining some insight into the issue they are facing and presenting the customer with a solution will pay dividends.
  2. Those who exited the flow prematurely for known reasons.
    • Sometimes APIs go down and the Viewed Under Maintenance Page is emitted. There could be latency in the network causing communications to slow down and Viewed High Latency Page is emitted. There are many reasons why these types of event can occur. A customer who is unsuccessful due to these issues is an excellent candidate for re-engagement.
    • The customer was unable to authenticate. A large percentage of end users exit the flow because they do not know the password for their payroll system or fall off during an MFA flow. This is another case with a high chance for success with a thoughtful re-engagement.
  3. Those who exit the flow prematurely for unknown reasons.
    • Sometimes life happens and you need to set the phone down and handle things. Sometimes your battery dies. If a customer initiates Transact and is in the middle of the flow when an onClose event is fired, re-engaging that user might be a good idea.

Next, we need to connect the dots between understanding where a user can fall out of the flow and the events emitted from the Transact SDK which can be used to identify these individuals.

Based on the groupings in the last section, we can consider processes with which to identify high-intent customers and automate reaching out to them.

  1. Multiple unsuccessful attempts:
    • For this case, you will need to collate error codes across users. In the data, look for customers who end the flow on Viewed Task Failed Page or Viewed High Latency Page events, for example, to indicate that the flow ended in an error state.
  2. Known error state:
    • The type of error will determine how to detect these users in the data and guide you to the type of re-engagement you should pursue. For example, users who fall out of the flow for authentication errors will encounter Viewed Authentication Failed Page or Viewed MFA Page.
    • There are also cases where the system was under maintenance or down for some reason. These are great candidates for re-engagement. Look for Viewed Under Maintenance Page. Once the system is back online, this user should be able to successfully complete the flow.
      You can hit the Company Search endpoint to determine if a company is currently in an operational or under-maintenance state.
  3. Unknown state:
    • This will be a catchall whereby you detect users who start the flow, but do not end in a success state such as Viewed Task Completed Page. Some of these will not necessarily indicate high intent. Some users will click into Transact, but have little intention of proceeding through the flow. We recommend looking for events which are triggered further in the flow to indicate a higher intent. Viewed Distribution Confirmation Page followed by an onClose for example. This user got most of the way through the flow, but did not complete for an unknown reason. Sending them a reminder to finish setting up their direct deposit may be a great idea.

The last step is to design the communications which will be sent to the customer to request they attempt the process one more time. Provide context for why now is a good time to try again via channels with which the customer is comfortable.

  • Write copy that entices your customer to attempt the flow again. Messages should be empathetic in nature. Present the customer with a solution to their specific problem.
  • Be sure to take timing of the communications into account as well. Sending out a message to a user on a holiday is unlikely to be successful.
  • You can use a variety of tools for actually re-engaging your customers. Text message, email, and in-app notification are all valid options. This will be highly dependent on the type of relationships you have with your customers.
  • As with any sort of marketing campaign, you will want to track the success of your re-engagement efforts. Utilizing a tool such as a CRM to track messages and responses is likely a good idea as these tools are custom built for this type of application.
  • To really drive success with your re-engagement processes, consider tying an incentive to entice your customers to try again.
  • Providing the correct context to your user can help to drive conversion. If they fall out of the flow because they do not have their password, remind them to grab it before starting the flow. You can also try out our new Conversion Opportunity webhook, if appropriate, to initialize Transact with the needed context.
  • Lastly, and perhaps most importantly, be sure to have a strong call to action that links back into Transact. Make it as easy as possible for your customers to be successful.

Setting up re-engagement processes is only one element of the overall value Atomic can unlock. Check out our other guides to get recommendations on building with the tools we provide.