The old model: reactive chat

The customer initiates the conversation
When chat options first became available on websites, they were based on the reactive model. A visitor to the site had a problem or a query, hit the chat button, and spoke to an agent who provided the necessary assistance or information. But the inadequacies and the missed opportunities of this model soon became apparent when the tracking and predictability of user activity became more advanced and sophisticated.
Limitations of reactive chat
Imagine you’re a store owner, and you see a customer walk in who shows an obvious interest in your merchandise, but you’re not allowed to speak to them unless they first approach you. You watch them walk out the door and think about how much you may have lost in sales through your enforced silence. That’s how reactive chat works. The visitor to the site must make the first approach, even though web analytics can easily show when they’re having difficulties and would benefit from being approached with an offer of assistance.
Superiority of the proactive chat model

With proactive chat, the visitor’s current and previous activity on the site serves as a guide to what they may need and when they should be approached with an offer of assistance or information. There are many different cues – called chat triggers – which indicate when the chat message should appear and what kind of advice or assistance should be suggested. Anyone who stays longer than 30 seconds on a site is usually at least a potential purchaser of what it has on offer. If they’re lingering without taking any further action, they may appreciate some assistance. Targeted proactive chat invitations maximize the chance of turning their visit into a sale.
Commercial benefits of proactive chat
The benefits of proactive chat include increased conversion rates, increased average order value, and a decline in shopping cart abandonment. A 2008 survey found that the average return on investment in proactive chat systems is 105%. Surveys have also found that online chat can create a 20% decrease in cart abandonment rates, along with a 20% increase in conversion rates and a 30% increase in the value of the average order. It also cuts down on customer care expenditure. Customers’ queries are dealt with directly as they arise, without the need to resort to the more expensive methods of phone and email.
Proactive chat triggers can help to increase a customer’s existing purchases by upselling, e.g.,
-“Hi, did you know that we’re offering a 20% discount if you buy 6 or more of product X?” or cross-selling, e.g.,
-“Did you know that product X gives 40% better results when used with product Y?”
Setting goals for chat
As with any site enhancement, the chat function has a dual goal – to retain customers and bring in new ones. More than three-quarters of businesses have found that holding on to customers is cheaper than acquiring new ones. Up to 70% of purchases from a business are made by 15% of its customers – that vital core who are most loyal to the brand. It costs five times as much to acquire a new customer as it does to retain an existing one. Therefore, a proactive chat function should be fine-tuned to the nuances of customer history on the site – how much was purchased, what kind of products, previous queries made by the customer in question, etc. Ideally, it should be designed to anticipate the future needs of the existing customer base.
The first step in designing a proactive chat function is to target the pages with a high exit rate. For example, if a majority of customers leave the page after 30 seconds or less, there is clearly a need to engage with them and offer assistance in the process of choosing a product, downloading an app, etc.
Determining what live chat trigger to use
Live chat triggers may be determined by what page they are on and for how long, what site (if any) they were referred from, whether they’ve visited the site before and how many times, their geographical location, whether they are on a mobile or desktop device, which products they are viewing, or their purchase history on the site. A previous record of the customer’s chat history should also be instantly available to the agent to assist with meeting their individual needs in the most efficient way possible.
The pattern of customer behavior indicates which type of assistance they would most appreciate. If they linger for a long time over the products page, it’s an indication that they may need assistance in deciding what to buy. If they fill their cart and then hesitate at the payment stage, it may mean they need some assistance in using the payment facility. Losing a customer at this point is particularly wasteful, and everything possible needs to be done to assist anyone who is clearly satisfied with the site’s products and is already in the process of purchasing.
Most common triggers
Hesitation trigger


Hesitation on the part of the customer is a prime chat trigger. Visitors who are not interested or who logged on by mistake generally leave within five seconds. Someone who stays but is making no further moves, therefore, may need advice or assistance in choosing a product, figuring out the pricing system, negotiating the payment facility, etc., depending on which part of the site they’re on. Typically, after a specific period has elapsed, the chat function will be activated with an automated message, e.g.,
-“Hi, would you like some help in choosing from our range?” and then the customer is connected to a live agent.
FAQ trigger


Someone who hesitates on the FAQ page, in particular, is clearly in need of information, and if their stay extends beyond a certain time, there’s a good chance that the answers on the page do not adequately meet their needs.
Location triggers


The geographical location of the user is a useful trigger. This can prompt a message in their native language and direct them to an agent who speaks it. It can also activate a message offering advice on shipping to the country in question or information on special offers for customers there.
Returning visitors


Return visitors can have messages welcoming them back, ideally tailored according to their visiting or purchase history on the site, e.g.,
-“Hello Katharine, do you have any questions about these new products we’ve added to our list?”
Referral trigger


If a customer has been referred from another platform with an offer on one of your products, this provides specific information on their intentions for the chat function to use, e.g.,
-“Hello, we see you’ve come about our special promotion on tablet devices…”
Seasonal and other triggers


The day of the week and time of year also provide excellent cues for chat triggers. Christmas, Thanksgiving, and Black Friday are all times to remind customers of special offers for the occasion. Thursday or Friday can be a good day to connect with customers who leave things to the last minute, e.g.,
-“Hi, would you like to know more about our weekend break offer? You can book now!” Even the time of day that a customer logs on can give clues to the type of information or assistance they might need.
Trigger buttons


Certain site buttons can be triggers in themselves, e.g., any link to a help or problem page on the site can be accompanied by an immediate offer of assistance. Making the “Contact Us” button a trigger also speeds things up by giving the customer the chance to get the information immediately in real-time.
Imminent exit


As mouse movement to the back button can also be detected, this is another opportunity to engage a customer who is about to leave the site, in case their exit is caused by something that might be solved by an agent, e.g.,
-“Hi, before you leave, is there anything we can help you with?”
Possible problems with proactive chat
No sooner had a proactive chat arrived on the scene than its critics were pointing out possible flaws. One obvious objection is that customers may find it annoying and pushy, especially if it’s not judiciously timed. Site managers must choose a point between activating chat too early (sounding pushy and intrusive) and leaving it too late (thus losing the customer). Setting the right time is crucial. But when properly used, proactive chat will more than make up in increased sales for what it may lose in customer irritation and withdrawal. More often than not, customer dissatisfaction is caused by a poorly-designed chat system, e.g., triggers appearing too soon or too often or the absence of an opt-out function that allows customers to switch them off permanently.
Ensuring customer satisfaction with proactive chat
Carefully calibrate timing
Timing triggers correctly is one of the most important parts of ensuring customer satisfaction. There are several criteria for the choice of time to offer customer assistance or information:
- total time spent on the site,
- time spent on the current page
- the total number of page views,
- number and duration of previous visits,
- the complexity of the page content
- the stage the customer has reached in the buying process.
Taking all of these into account can greatly enhance the customer’s reaction to proactive chat and turn it from an irritating to a helpful and satisfying experience.
Avoid spamming
To minimize any negative feelings about the chat, keep the invites to a minimum. Customers don’t want invites popping up on each page they visit. One survey showed that too many chat invites were among the primary reasons customers left a website.
Customized invites
Matching invitations as closely as possible to the customer’s likely needs is another way to improve their satisfaction with the feature and reduce the risk of a negative response. Whether they are a new or repeat customer, the page the visitor is on, a promotion link that brought them to the page, can all be used to customize the chat invite they see.
Keep it simple
Customers need to feel in control, i.e., they should be able easily to accept or decline a chat invite and not have too much difficulty in using it if they choose to. Providing the simple options, “Yes, I’d like assistance” or “No, I don’t need any help right now”, can make a lot of difference to the customer experience. When they do avail themselves of the invite, it should be simple and involve as few steps as possible. Minimize the interruption to the customer’s shopping experience. Their primary purpose on the site is to buy, and any other activity that slows down that process is damaging.
Opt-out function
Also, a site needs to have an opt-out from chat if the customer so chooses, a choice that should be remembered on every subsequent visit to the site, without their having to click “No” each time they log in. In this way, any customer who does not want any chat prompts can be free of them for good, while retaining the option to ask for assistance if they need it.
Targeted offers
Another way to improve the experience for customers is to make offers targeted to their activity. Surveys show that customers who have not chatted in the past are more open to doing so when a special offer is included. This can be customized to match the precise point the customer has reached in the process of choosing or purchasing a product. For instance, if they have filled their shopping cart but are hesitating before checking out, they can be introduced to a special offer of a reduction on one of the items they have selected. This can encourage the customer to proceed to the payment stage and complete a sale that might otherwise have been abandoned. They can also be reminded when they’re getting close to the free shipping threshold in case they want to buy more and eliminate shipping costs.
Following up with good service
Rather than consider the possible negative effects of chat, it makes more sense to talk about both the live chat triggers and the quality of the information/advice combined. Neither can be seen in isolation. A trigger may be irritating if followed by poor quality chat service, while the customer may welcome the same trigger if the advice or information is genuinely helpful. Nothing is more annoying than to put a question to an agent and be answered with a vague and generic response that was obviously picked from a prepared list. Anyone planning to use proactive chat, therefore, must make sure to have a first-class system of customer care in place to begin with.
Conclusion
As with most useful innovations, proactive chat is switching from being an optional extra to a necessity. Businesses cannot compete without it. It is not a matter of whether to use it, but how to design it to best effect for the site in question, and how to adopt best practices proactive chat tailored to its exact needs. Site statistics need to be constantly reviewed to ensure the chat function is working at optimal efficiency. In the future, it won’t be a question of using proactive chat or not, but how to design the best system for your site.