Important Chatbot KPIs That Every Marketer Needs to Stay On Top Of
It's not uncommon to hear that you can't adjust that which you aren't measuring. Here are the most important chatbot KPIs that you should track at all times.
Congratulations! You have finally implemented a chatbot into the customer service strategy of your medical practice.
You're ready to pour yourself a drink, kick up your feet, and let the chatbot do all of the work.
While we highly encourage the former, a medical chatbot should never be left to its own devices without making sure it's doing the work it's intended to.
A KPI, or, Key Performance Indicator, is used to ensure that your organization is achieving its key business objectives with the tools and strategy it has implemented. To learn more about the most useful chatbot KPIs, check out our guide!
Best Chatbot KPIs for Medical Chatbot
Why are Key Performance Indicators so important?
Through daily, weekly, or quarterly reports, business KPIs allow an organization to:
Monitor the long-term and short-term health of a company
To view annual and quarterly progress
To know where to make adjustments
To see opportunities for improvement
To see patterns developing over time
We can then take these considerations and customize our chatbot analytics to bring visibility to specific performance metrics. These are some of our favorites for the medical and surgical aesthetics industry.
Total Users and Total Conversations
The number of patients using your chatbot is important because you'll be able to see how many site visitors have an interest in your support tool. The total user metric will indicate how much user data your chatbot has gathered, and the assumed market size.
Conversations (new or total) will capture both new customers and those returning. As a KPI, it can also break down how the number of initiated conversations compared against the conversations completed by the chatbot.
It's not enough for a chatbot conversation to be initiated. The "active users" metric is crucial, as it will demonstrate how many people are actually reading the content in a clearly defined time frame. Even if they don't engage, they are reading your content.
Engaged users are the patients who are communicating with the chatbot. They are both receiving and sending messages, and likely believe that the chatbot is a useful resource.
In order for a chatbot to learn, it must receive data from a user session. If the chatbot has enough "buy-in" from its users, it usually indicates that the chatbot is high-quality and that it's meeting the needs of the patient. They'll continue to return when they have questions or scheduling concerns.
Goal Completion Rate (GCR)
A goal completion rate represents the percentage of successful interactions through chatbot engagement. If a user is trying to find a specific answer or get more information on the types of services you provide, the chatbot's ability to receive the inquiry and provide the requested information will be a huge marker of success.
Analytics will show the user input, path, and outcome. Due to machine learning, the more often a goal is completed, the more likely it will be successfully completed in the future.
Many chatbots are designed to offer an exit survey once a request has been resolved. Patients using your chatbot will be able to indicate if their question was successfully resolved and if the chatbot communicated effectively.
User satisfaction can be measured by answering a simple "yes or no" binary survey (did the chatbot perform well?), or as complex as responding to a multi-point NPS (Net promoter score) survey.
Goal Completion Time
Chatbots are only efficient if they can resolve your customer's request in a reasonable amount of time. A seamless and efficient experience will ensure that your patients return to the chatbot for help. Tracking the amount of time it takes for a chatbot to resolve a request (or the number of messages/taps) will help you visualize where the chatbot needs additional programming.
Your medical chatbot will do more for you than just connect with your customers. Ideally, it will be hard at work collecting customer data and generating leads for your practice.
A chatbot can influence a patient's interest in a service or a particular provider, and make scheduling a seamless task. Additionally, the data collected by the initial interaction can then be used in alternate marketing channels to promote your services.
Fallback or Failure Rate
You might think that running a chatbot will ensure you avoid human error. While the human aspect is somewhat removed, chatbots should still be expected to fail-- especially during the beginning stages of machine learning.
A "fallback" or failure rate is the measured percentage a chatbot fails or experiences near-failure.
Failure can be measured in one of three ways:
Confusion rate - chatbot does not understand the user's question
Reset rate - the number of times a user wants to change their previous response/refreshes
Human takeover rate - the percentage of users who request a transfer to a live agent
Once the chatbot has time to learn and respond to patient inquiries, this fallback or failure rate should improve.
Your practice will need to consider how much it is spending on the chatbot, and how much business its implementation is generating for your business. During this process, your practice should consider:
Cost per fallback
Once you have this data, you can see how the ROI performs compared to other communication or lead generation channels your marketing team is currently implementing.
Improve Your Chatbot Performance
A chatbot should never be a tool that you just "set and forget." Chatbot analytics will give you the clearest insight into your chatbot KPIs, so you can make sure you are serving your patients to the best of your ability.
Interested in learning how the AesthetiBot works? Check out our helpful guide into how implementing our chatbot can meet your patient's needs.