This is one in a series of Anti-Churn Tactic articles detailing ways to reduce churn and improve customer retention. In this installment, we'll talk about using in-app chat to help users succeed.

For this series, let's assume we have an invoicing application named InvoiceViz and customers are signing up like mad because the app is beautiful and the prices enticing. But we're finding that those users aren't sticking around as long as we'd like. Something has to be done about churn.

The first thing we should do is find out why they're leaving and focus on any actionable feedback. That much goes without saying. But let's say we've done that and now are left with otherwise happy users who just need a little prodding and support to stay.

In-app chat is a great way to snag users at the moment they need you.

We've all seen in-app chat boxes. They are those boxes in the lower right corner that pop up in the middle of, say, buying underwear online with a friendly operator who types "Hello, is there anything I can do to help you with your purchase?"

In-app chat boxes are a great, cost-effective way to connect users to real humans at exactly the point when the user might need some extra support. Because they are semi-asynchronous and not all users engage with them, a single customer service agent can serve many users at the same time.

Lots of tools exist to help with in-app chat.

We've seen clients use several different tools. They all work basically the same way so it's largely a matter of preference. If we had to name a few we'd say Olark, LivePerson and Moxie LiveChat. But really, what matters most is the results we get. For example, we had one client gain valuable insight about rental book returns after just a week or two of slapping an Olark chat window in the middle of the process.

First, try not to be creepy or overbearing.

Following users around with a chat box on every page saying "hey do you need help?" or "we notice you areā€¦" or "I'm just right here if you needed meeee..." is kind of like that annoyingly over-eager sales person who follows you around a clothing store commenting about every outfit. Users want help, but they also want to be left alone long enough to think. We need to strike a delicate balance.

Drive chat window placement with data.

To the extent possible, we'll gather data on the places where users are stumbling the most. Have a 3-step workflow where 20% of users drop off between steps 2 and 3? Put the chat window on step 2. Have a form that creates an unusual number of validation errors with user input? Bam, chat window. Have a general-purpose home page with 5 different, unrelated calls to action? No, no chat window for us. We'll clean up our homepage instead.

Make sure your chat window is well staffed.

Letting customers chat with real humans is great. But it can backfire if that chat experience is frustrating, And there's nothing more frustrating than a stilted, delayed, distracted chat conversation with an overworked customer service agent. Or struggling with an ill-trained customer service agent who wastes the user's precious time. So, before deploying the window, we must make sure the following basics are covered:

Revisit placement choices periodically.

Chat windows are a high-touch, high-cost way to improve customer retention. They should not be a permanent fix for anything. What they should do, instead, is give us insight into difficult or confusing parts of the application. And this insight should lead to changes. And change should lead to reduced confusion. Eventually, you'll hone the application and process to a point of diminishing returns. This is the point at which its time to remove the chat window (and the customer service burden that goes with it). Then, redeploy those customer service resources to some other hotter hot spot in the same application. Oh, and did I mention that finding those application hot spots is what we do here at Pollen?

Like this article? Read the next in the series about using discounts and coupons to reduce churn.