5 Practical Ways to Automate Ecommerce Customer Service
Consumers are buying online in greater volumes than ever. Sales are up, which is great for business, but it also puts extraordinary pressure on customer service and fulfillment teams who are supporting more orders with the same (or even fewer) resources.
This pressure increases the chance of mistakes while providing service, which can lead to reduced customer satisfaction and lost future business. Fortunately, ecommerce businesses can use the power of help desk automation to improve their customers’ buying experiences by flagging and prioritizing customer issues in real time.
This article will cover five of the most common ecommerce customer service automation options, but there are many more. Spend a few hours reading what your customers are saying in support tickets or chats — or consider using a customer conversations analytics tool to gather insights.
1. Prevent cancelled orders from being fulfilled
It happens too often: As soon as the customer creates a new order, fulfillment starts and the package is put together by the warehouse team. If the package is ready immediately before the courier collection time, it is handed over to the shipping company right away.
In the meantime, the customer has changed their mind or realized they’ve made a mistake. They reply to the order confirmation email in order to cancel it, but by the time your next available agent reads their message (which may be hours later), it's too late.
Now the only option is for the customer to wait for the package to arrive and then return it, which is costly for the business and inconvenient for the customer.
To prevent this issue, configure your help desk to identify incoming cancellations and trigger a useful action. That action may simply mark the conversation as urgent or automatically pause/cancel the fulfillment of an order.
The latter option would use a webhook configured in your help desk and, depending on your ecommerce platform, might require some development resources.
To identify those cancellations, a simple option is to have a drop-down field on your contact form allowing the customer to self-select their reason for contact. To capture incoming cancellations from other channels, use workflows to look for common words and phrases. A quick read through historical order cancellation emails will help you find examples of the right words and phrasing.
At Cx Moments, our ecommerce customer musicMagpie uses auto-tagging in just this way, saving related delivery and return costs, but most importantly, improving their customer’s experience.
2. Give frustrated customers the white glove treatment
If you aren’t careful, unhappy customers can leave after a single problem and never return. In addition, they can make their frustration public in just a few clicks, with long lasting damage to your brand and credibility.
Frustrated customers tend to use similar words and phrases to express their feelings about an issue with their purchase. Whatever language your customers use, online dictionaries will provide a helpful starting point for identifying the most likely trigger words.
In addition, reading through actual customer conversations and noting the words customers use when frustrated can help expand your list. That list of words and phrases can then be used in rules to identify the most upset customers and handle them more effectively.
For example, you may assign a higher priority to these cases so the customer gets a response more quickly, or you can route them to specific agents who are more skilled in managing difficult customer interactions.
3. Prioritize incoming requests
Answering your customers strictly in the order which they asked for help feels fair, but it isn’t necessarily the most effective option. Not all queries are equally important.
In ecommerce, post-sale/pre-shipping inquiries such as address correction or order modification are good examples of high-priority issues. If your customer has made a mistake in their delivery address and the package gets shipped before it can be corrected, this will negatively impact both the customer experience and the cost to your business.
Pre-sale problems at the checkout stage of the buying journey are also critical: Difficulties with adding products to a basket, applying discount codes, or making payments are all moments where a customer who cannot get quick help is likely to move on to a competitor, and they may never return.
You do not want those questions languishing in your support queue while your team is tied up dealing with less urgent questions like collaboration and partnership requests, job applications, or email unsubscribe requests.
Read through past conversations in your help desk to identify your highest and lowest priority question types. Use tagging and workflows to change the priority level of those conversations, and give your team that context when deciding which ones to answer first.
4. Spot damaged items
There are few things more destructive to the customer experience than paying and waiting days or weeks for a defective product. Not only will it create additional costs (in packing, shipping, and returning), but more importantly, it will impact your reputation and brand.
While you can’t completely avoid damaged products, you can certainly spot patterns early and try to correct them.
For example, there might be one product category or product type that is more likely to be defective or damaged in transit. These patterns can be quickly detected if the team responsible for procuring or building the products are notified every time a new defect is reported by customers.
One simple way to let the product team know when a customer complains about this issue is to automatically tag any conversation that mentions a defective product and use business rules to trigger an email notification or @mention to the right person or group.
In Help Scout, for example, this can be done using workflow conditions and actions, but any modern help desk should have a similar capability.
5. Notify your carriers of delivery-related complaints
No matter who your shipping provider is, there will always be issues with delivery: Customers not receiving the order or receiving it too late, misplaced parcels, driver behavior, and so on.
While this is a part of your customer’s journey that is largely outside your direct control, setting clear expectations and effective communication with your shipping partner will definitely help.
In the same way you can notify internal teams about product issues, you can notify your external partners, too. With some custom configuration, you could add additional information, like the tracking number, directly into the email to help the carrier quickly investigate the issue.
By the time your agent is handling the conversation, they might have the response from the carrier already available. Ticket resolution time and time spent by agents will both improve, with positive impacts on customer experience.
The future of help desk automation
The automation examples described above can be implemented in most modern help desks that provide a way to configure and trigger business rules based on a list of words found in the issue description.
However, depending on the kind of language your customers use, you might find there are many different ways a customer can describe the same problem.
This is one of the features of human language and a challenge that AI developers are trying to solve using machine learning applied to customer service. Rather than using keyword rules, an AI algorithm looks for patterns (or commonalities) in past examples to understand what the customer is asking for.
As a rule of thumb, use simple rules whenever the words your customers use to describe an issue are unambiguously linked to that issue and rarely mentioned in other contexts.
For example, when customers are complaining about damaged products, they may use words such as "damage," “defect,” or “flaw,” which are less likely to be used in other situations. An issue description containing any of these words is probably referring to a product being damaged (in the ecommerce context).
On the other hand, a word like "stop" can be used in different contexts: “Please stop this item from being delivered” (cancel my order) or “Please stop ignoring my email” (respond more quickly). Using the word “stop” as a simple rule to detect cancellation requests would trigger too many false positives.
In this case, an AI-driven approach may be more effective at automatically categorizing the issue correctly.
Customer service automation, when done in a customer-centric manner, can bring great benefits to your customers and your team without losing your brand voice or human touch.