How to Prepare for the Next Major Digital Touchpoint, Conversational Commerce
Conversational commerce means interacting with your customer in an automated dialog via voice or text.
Usage of voice-based conversational interfaces such as Alexa and Siri are exploding. Meanwhile over 100,000 active bots have been created on FaceBook's messenger platform as it approaches its first birthday.
"Digital" began to truly scale with the web, then expanded even further via apps and social. Is conversational commerce (CC) the next major touchpoint? Conversational commerce is defined as interacting with your customer in an automated dialog via voice or text. Based on our experience consulting large brands on the implementation of their digital innovations, and given trends on consumer adoption and technology-readiness, it is fair to say that we are at the dawn of the first wave of the broad implementation of conversational commerce.
For several years, IBM has been painting a fanciful picture of its Watson technology's ability to digest volumes of information, understand questions, and provide truly insightful answers. Meanwhile, consumers are becoming more and more comfortable with dialog-style interactions via Siri, Google voice search, and Alexa.
But conversational commerce does not necessarily have to involve voice recognition.
We are now at the one-year anniversary of the launch of Facebook's chatbots, which enable brands to engage in text-based automated interactions with their customers and audience. Its Messenger platform alone has about 100,000 active bots, and a recent survey found that nearly 80 percent of companies use or plan to use chatbots by 2020.
At the Shoptalk conference last month, eBay President and CEO Devin Wenig announced the launch of eBay's new chatbot called ShopBot, which advises customers on items they might like to buy via automated chat dialog.
This pattern makes sense, as we see that millennials - 38 percent of whom prefer texting as their number one form of interaction, according to a study from Think with Google - have elevated this type of communication to an art form.
Will Siri or SMS-like automated dialog with your brand become the next big consumer touchpoint? If so, what do you need to do to prepare?
The answer, as some of the stats above suggest, is that conversational commerce is poised to be a major and preferred interaction model for many future brand interactions.
The good news is that if your brand has built a reasonably flexible and integrated digital stack, it can often be quickly leveraged to enable high-value CC capabilities without requiring that you install "Deep Blue" in your data center.
Here are five key things to know about getting started with conversational commerce:
1. The core of conversational commerce is very similar to search.
If you already have a strong search platform that permits parametrization, you can use it to drive a key portion of your chat experience. When you tell eBay's chatbot you are looking to buy a voice recorder, it asks you questions such as the size and memory capacity you need. These questions are simply the metadata parameters eBay has available for voice recorders. You can utilize the product metadata in your existing catalog to make your chatbot appear to ask smart questions, and even more importantly help the customer find what they need, but in reality the results are very similar to what they would experience if they simply entered structured search queries.
Of course not all queries involve the quest for a product. Some may be asking a question, such as about your return or cancellation policies, but this too is very similar to search. You can parse chat questions against your full text index and return intelligent answers by, again, leveraging your search engine.
2. The next step of conversational commerce is about enabling transactions.
Once a customer has found what they are looking for, they may wish to buy, reserve, add to a wish list, or take some other action. Your chat flow needs to know when to pivot from searching to asking the customer to take action. In many cases, or in your initial releases, you may simply choose to branch to existing web screens to complete transactions, as eBay is doing with ShopBot. More sophisticated conversational commerce implementations allow the customer to take action via voice or text, such as Domino's, which allows the customer to order a pizza by text.
3. In text conversations, you generally know your customer.
One of the advantages of most forms of conversational commerce, such as SMS or Facebook Messenger, is that your customer is identifiable. If the customer has a profile in your system, you can use this knowledge to make the conversational interaction simpler—and also smarter. Picking up again on our Domino’s example, when the customer texts them a pizza emoji, the bot matches their telephone number to its database and confirms that it will be placing the order with toppings based on their past preferences, and will deliver it to their home address on file. The customer will then have the opportunity to override any of these defaults if they are in the mood for Hawaiian pizza that day.
4. You can make the language parsing easier by giving multiple-choice options.
Many successful chatbots are more of a string of multiple-choice questions than a free-form dialog. This substantially reduces the challenge of “comprehending” the customer and furthermore reduces typing for users on mobile devices. Naturally you will want to support customer-entered text strings, but a considerable number of interactions can be handled via a series of multiple-choice questions. In some ways, chat is similar to IVR systems at call centers, and can often use similar types of decision trees.
5. It doesn’t have to be perfect.
There is still some novelty to automated interactions, so customers don’t expect them to be perfect. Furthermore, as with any digital platform, you have the opportunity to iteratively improve it over time. Siri has grown tremendously over the last few years in the range of queries it can handle.
A fantastic resource to help guide your prioritization of new capabilities are the chat logs themselves, which will give you a sense of the types of interactions that your customers are attempting that may not yet be supported by your platform. And in the meantime, as you become aware of such chat or voice requests, you can create short text responses to those categories of inquiries, letting the customer know what other touchpoints currently support that action. So if a customer, for example, uses a chatbot to check their account balance but then wants to transfer funds, and that is not yet supported via conversational commerce, you can supply the URL for the website or app and the toll-free number to call, so they know where to go next.
6. Develop a core conversational engine, and leverage it across many different touchpoints.
It makes sense to invest in conversational commerce platforms and tie them to your existing catalog, customer data, business logic and transaction capabilities. In doing so, think of creating one central CC "engine" that will connect to a variety of conversational endpoints. To begin with, you may want to focus on enabling a chatbot on your website(s) and in your apps, and integrating with the Facebook chatbot API to allow customers to chat with your automated system via Messenger the same way they would chat with their friends. But in future iterations, it makes sense to support SMS, Skype, WeChat (if you do business in Asia), and possibly other similar platforms. Longer term, as Apple's Siri, Google Voice, Microsoft's Cortana, and Amazon's Alexa continue to open up their APIs, the same conversational engine you created for text can be leveraged with relatively small modifications to support voice interactions.
Conversational commerce is already here and most major brands have either implemented or are in some stage of planning around an implementation. You can probably leverage existing systems and data sets to create a reasonable starting point for conversational interaction without requiring sophisticated AI or language parsing. Over time, you can learn from your customers’ queries how they want to interact with you and evolve your conversational capabilities accordingly.