A How-to Guide on Building Smooth Chatbot Experiences for Ecommerce Platforms

David Stepaniuk

Aug 24, 2021 • 11 min read
chatbot

Chatbots have grown in popularity over recent years, as they can be great tools for engaging website visitors and app users, improving their experience and, ultimately, achieving higher conversion rates. But there’s a dark side to this ubiquitous technology. Chatbots, when deployed incorrectly, can be disruptive, even to the point of chasing users away. Let’s take a look at the steps you can take to avoid this outcome.

Chatbots - what they are great at

The number of use cases in which chatbots excel is impressive. Let’s take a look at a few examples with data to back them up.

Chatbots can be used to drive sales in various contexts, including ecommerce platforms and service providers.

How do consumers use chatbots? Mostly to find a quick answer to a question. The flipside of this is that users expect chatbots to have the answers and to offer them around the clock. 41.3% of shoppers make purchases through chatbots and 33% would happily make a reservation at a hotel or restaurant through a conversational tool. Overall, users seem to enjoy conversing with chatbots, which boast response rates of 35-40%.

Meanwhile, companies can save 30% on customer support and generate more high-quality leads. Ecommerce businesses have begun looking into more specialized use cases. For example, using abandoned cart chatbots and private messaging on Messenger in tandem can lead to a revenue increase of 7-25%. Chatbots offer complex support for ecommerce, from lead generation, through handling the entire sales process, including payments.

To sum up, chatbots can serve as a great tool for:

  • Driving sales,
  • Improving customer service,
  • Generating more and better leads,
  • Boosting customer engagement,
  • Building a positive brand image.

When do chatbots fail?

For the most part, customers have positive experiences with bot-only chats – their average satisfaction rate is 87.58%. But few things are more frustrating than a chatbot that interferes with the shopping experience or doesn’t help you get a desired outcome. When consumers hesitate to use chatbots, it’s mainly because they expect to be misunderstood, or that the chatbot won’t be able to solve their problem.

Both chatbots that pretend to be human and chatbots that are open about what they are can introduce problems. When chatbots introduce themselves as chatbots, purchase rates go down by 79.7% and more. Undisclosed chatbots tend to be more effective sellers than inexperienced human workers. However, chatbots that try to imitate sophisticated human interaction and fail can have a strong negative effect on the brand. The majority of users (61%) get frustrated with chatbots more easily than with human interlocutors, particularly when a chatbot doesn’t solve their issue.

As such, it’s better that the chatbot passes the baton to human customer service, or a sales representative, for more complex issues. A badly-scripted, inappropriately applied chatbot might seem like a cheap replacement for real customer service that only prevents users from reaching a human they want to talk to.

Sensitive information is another major challenge. These days, most people are aware of the risk of identity theft and online fraud. It’s likely the reason why almost 43% of users would rather share their data with a human on the phone than a chatbot. This issue, at least, could be solved through improved cybersecurity measures and good communication with users.

Finally, technical issues like a lack of specialized tech talent to take care of implementation and platform fragmentation (when a company needs a separate chatbot solution for their Android, iOS, and web apps, for example), may stop businesses from investing in conversational marketing tools.

A how-to guide on building a smooth chatbot experience in ecommerce platform

Chatbots are detrimental to business in several cases:

  • When they don’t solve users’ issues,
  • When they are not helpful in the context they are placed in,
  • When users don’t trust them,
  • When they disrupt the experience a user has within a specific platform.

All of this points not at technology, but implementation as the culprit. Is it always possible to use a chatbot successfully in a given use case? Not necessarily - sometimes it’s truly better to choose other tools. But there is a way to make sure that implementing a chatbot is the right choice for your business and that the chatbot will help users, not hinder them.

Here’s our step-by-step guide that will help you do so.

Step 1: Apply ecosystem mapping or a service blueprint

The first step to crafting a successful chatbot experience is to discover the weak points in your platform’s UX. These gaps where people need help are perfect spots for a chatbot to come to the rescue. How to find them? Ecosystem mapping and the service blueprint are two helpful approaches.

Ecosystem mapping focuses on visualizing how a product or platform works: what systems, processes, and data flows it includes, and how a user moves around it. An ecosystem map will help you understand each element that’s part of your platform and all key interactions that affect user experience. Once you have this map, you can plan where and how to add a new element to the ecosystem - in this case, a chatbot.

To prepare a useful ecosystem map, start by drawing the user at its very center.

Next, add key people and groups that the user is connected to (such as a sales representative). Draw the channels and tools that allow for communication between them (smartphones, desktop computers, etc.). Finally, fill the map in with data flows, monetary transactions, and any other interactions specific to your business. The finished map will help you see where a chatbot might help you achieve business goals.

A service blueprint is a diagram used to show the flow of a specific user journey, with a focus on its various components (people, tools and processes) and how they interact with each other. They are a great way to visualize omnichannel experiences with multiple touchpoints. The scope of the blueprint depends on the user journey and the user goals achieved within it.

A service blueprint can help you find weaknesses within each journey, which in turn creates opportunities for optimization - such as through introducing a chatbot. Here’s what it may look like.

Step 2: Define challenge areas for chatbots

Chatbots are tools to achieve specific goals. That’s why looking for user pain points and opportunities to improve your platform’s UX is a great way to discover areas where a chatbot might come in handy. For example, if users of an ecommerce store have questions about a product, where can they find relevant information? Would their inquiries be too complex and varied for a chatbot to handle (e.g. for custom-made products)? Finding the right space for a chatbot to shine requires a lot of thought.

Step 3: Define target users and the value proposition

Before you deploy a chatbot, you need to know who it’ll be talking to. That will allow you to construct effective conversation scripts.

Additionally, make sure that the chatbot’s value proposition is clear - both the value that it will deliver to users (such as quick access to information), and the value it will generate for your business (such as a higher conversion rate).

Step 4: Use 2 filters - business and user outcome + technological enablement

A great way to find an area of opportunity that’s just waiting to be filled by a clever solution, such as a chatbot, is to look for the overlap of business and user outcomes and technological enablement. The important part is choosing solutions that have the best chance of producing desired outcomes.

At Netguru, we use workshops to sift through competing ideas and solutions. Two logical filters help us focus on efforts that are feasible and likely to create the greatest impact for our clients:

  • Filter 1: Business & user outcomes - we look for ideas that best support customer and business needs,
  • Filter 2: Technological enablement - we consider current technological ambitions and constraints.

Step 5: Begin the iterative development process

It’s important to always be ready to learn from your mistakes or change your approach based on new information. That’s why the iterative approach is so valuable - it keeps you flexible and helps you adapt to the needs of your users and business.

The process begins with initial planning, but assumes that plans might change. Each iteration (often a Sprint, using Agile and Scrum terminology) is planned individually, using current requirements, data, and design insights. Then, implementation happens, followed by testing and evaluation of outcomes. It’s a structured way of ensuring that the project continues heading in the right direction.

Step 6: Measure, control and improve

Based on an iterative approach, you’ll be able to use trackable metrics to make strategic choices about your chatbot implementation. By establishing KPIs that help address business and user needs and analyzing the gathered data, you should have a much easier time noticing opportunities for improvement and measuring your progress. A big part of this is a good quality assurance process, one that’s focused on your goals and optimized for top efficiency.

Key takeaways

The important thing is to remember that not all users approach chatbots the same way and that each specific use case should be approached individually. Chatbots can bring in tremendous gains, such as higher user satisfaction and improved conversion rates - but they need to be placed well within your business ecosystem.

  • Most users interact with chatbots to find answers to specific questions - quickly.
  • Over 40% of shoppers make purchases through chatbots, and this number may grow.
  • Companies can save 30% of customer support costs and increase their revenue by up to 25% by deploying chatbots.

To reap these benefits, make sure to map out the components and interactions crucial to your business, find the right problems for chatbots to solve, address your users’ unique expectations, and choose a development team that works iteratively based on solid metrics.

More posts by this author

David Stepaniuk

Disruption Forum Retail 2021