Mission: increase conversion rate to starting a dialog in a chat

A live chat is not only your ally in supporting users but also an invaluable source of leads. Despite the fact that this tool is often underestimated, a chat will help convert interested website visitors into real customers and lead them to the purchase.

Considering a chat in these circumstances we can easily conclude that the more users contact us, the more potential customers we have. All you have to do is to make sure more users contact you via chat. We’ve talked a lot about proactive trigger messages that can help you with it, which you can send to particular user segments at the right time and involve them in a conversation. But that was not enough for us. We were thinking, “how do we get even more attention to a chat?”

Let us remind you that the “classic” version of a chat message in Dashly looks like this:

The easiest way to gain attention is to make the message bigger. This is exactly what we did: we added the agent’s name and a userpic making the message as similar as possible to a chat in social media so that the agent is perceived as a person and a friend.

Then we realized that it is not enough to attract users’ attention, we have to “make” them answer something. To do this, we’ve put the answer box right below the message so that the user can text us right away.

We made an experiment: we took all three options and brought them up against each other not just in the A/B test, but in the A/B/C test!

Read also: Capture emails from website visitors with Pop-ups

For our experiment, we have chosen 4 online services from among our customers. We took several websites on the same subject for two reasons:

1. their total traffic will help achieve statistical significance within a short time;
2. averaged data from several services eliminates errors caused by website or traffic specificities.

We wanted to prove that a chat message with an input field will increase the conversion rate to beginning a dialog by at least 1.5 times.

To test this hypothesis, we launched an automated message with the simple text “Hello! Can we help you with the selection of your plan?” that appeared 20 seconds after a potential client entered a website. The message was shown once and deleted after a minute if the user did not reply to it.

Read also:
⭐ How to find respondents for a survey or customer interview
⭐ 2025 Email Sales Funnel Strategies: How to Convert More Leads [11 Examples]
⭐ 6 Smart Tactics to Solve Your Sales Funnel Leakage Problem

The method of dividing users into groups is simple: we randomly assigned a number from one to three to website visitors and depending on the number, a user saw the respective message. When they returned to the website, the new version was not shown to them so that the test results were not mixed up.

The experiment lasted for 3.5 weeks (from December 16, 2019, to January 9, 2020); the total traffic was 130,185 unique visitors.

The conversion rate to starting a chat often seems small, but in absolute numbers, it’s more than a hundred potential buyers in each experimental group. Pay attention not only to the percentage but also to the number of leads (and ideally purchases) to see if the chat expenses are paying off.

We’ve calculated the conversion rate of each service individually, and then the average rate on the total traffic. As deviations are insignificant, we can accept that the situation in all four online services is more or less similar (and we can extrapolate this situation to other services).

Read also: The North Star Metric and How Can This Lead to Success

Use what’s best

We surely want you to have the best and most efficient conversion ways to communicate with your leads. That’s why we’ve replaced the appearance of chat messages: we added a sender’s userpic, the time, and a response input field. This is what any chat messages look like: triggered, messages from your agents, and automated replies.

Read also: Find out the best Intercom alternatives and Drift alternative

How to make it even better

We’re intentionally talking about leads in this case and not sales. When an agent gets involved, Dashly, as a service, cannot influence the conversion as the human factor is too influential. We already have some ideas on how to help you deal with this, but in the meantime, here are some recommendations that will help you work even better with those who have contacted you via the chat:

• Understand the task. We showed the message to everyone to get the maximum reach. Think about it; maybe you only want those who have a high order value in their cart or those who are interested in new products to contact support. By segmenting the audience, you can pick a better offer and convert more users, but in absolute values, the response rate is likely to be lower.
• Add a userpic and the agent’s name. This is extremely banal, but still: people better respond to people. In this experiment, the automated message was sent from Polly with her photo so that different names of agents would not affect the test results.
• Personalize. You can insert the information that you already know about users into messages. For example, their name, products in their cart, or you can build your offer based on what page they visited. The text is changed automatically, the message goes straight to the user’s heart, and your conversion rate increases.
• Work with the tech support team. It’s not enough to just acquire a customer, it is important to communicate with them properly so that this communication turns into a purchase. Strange as it may seem, not all support teams are ready for an inflow of inquiries. Read our article on how to optimize the work of the support team and enable dialog analytics to better assess the situation.

Have nice and successful conversations!