An AI chatbot is an investment, and like any investment, you should measure the return on investment (ROI). Fortunately, a chatbot's ROI is easier to measure than that of many marketing and technology purchases, where it can be hard to tell which result came from which tool. The effect of a customer service chatbot can be tracked with concrete numbers.
This guide covers the main measures to track, how to calculate ROI, and the targets you can reasonably aim for.
The Four Sources of Chatbot ROI
Chatbot ROI comes from four main sources. Understanding each one helps you see the full value.
1. Cost Reduction
The simplest part of ROI is lower support costs. Every conversation the chatbot resolves without staff involvement saves staff time. To calculate this, you need your cost per ticket (support request), which is staff salary, benefits, tools, and overhead divided by the number of tickets handled.
If your cost per ticket is 15 dollars and your chatbot resolves 500 conversations per month that would have been tickets, that is 7,500 dollars in monthly savings. Compare that figure with what you pay for the chatbot each month to see the return.
2. Revenue Generation
Where a lead goes after the chat
- A visitor opens the chatThe widget appears on every page where you pasted the snippet.
- The bot asks your questionsOne at a time, with buttons for the common answers.
- Alma saves the leadIn your Leads inbox, with tags from the visitor's answers.
- You get an emailAs soon as the visitor leaves an email address or phone number.
Where else the lead can go
- Every plan
- Email notification
- Growth and up
- Zapier, CampaignCNX+ and CoreCNX
- Agency and up
- Webhooks to your own systems
Chatbots that collect leads (contact details from interested visitors) and help with sales produce direct revenue. Track how many leads the chatbot collects and follow them through to a sale to see how much revenue came from the chatbot. Even if the chatbot mainly answers support questions, answering questions before a purchase helps visitors decide to buy, and that effect can be measured.
3. Customer Retention
Faster responses and 24/7 availability can reduce churn, which is the rate at which customers stop doing business with you. This is harder to measure directly, but you can compare churn rates before and after adding the chatbot, and between customers who use the chatbot and those who do not. Even a small reduction in churn can add up to significant revenue over the life of each customer relationship.
4. Operational Efficiency
Beyond direct cost savings, chatbots help your team work more efficiently. Support staff handle fewer routine tickets and can focus on complex issues. This often leads to higher staff satisfaction, lower turnover, and better results on the tickets that do reach a person.
For a broader overview, read measuring whether a chatbot is working.
Key Metrics to Track
Deflection Rate
This is the percentage of customer conversations the chatbot resolves without staff involvement, and it is your main measure of efficiency. Calculate it as chatbot-resolved conversations divided by total customer conversations, multiplied by 100.
What to aim for: Expect a modest deflection rate in the first month, rising steadily over time as you improve the chatbot's knowledge base (the business information it answers from).
Resolution Rate
This measure shows how many of the chatbot's conversations are fully resolved and how many need follow-up. A high deflection rate is of little value if customers have to contact you again about the same issue, so track how often customers come back after a chatbot-resolved conversation.
What to aim for: A large majority of chatbot-handled conversations should be fully resolved.
Average Handle Time
Compare how long the chatbot takes to resolve a conversation with how long a staff member takes. Chatbots usually answer common questions within seconds, while staff members often need several minutes for the same types of questions.
Customer Satisfaction Score
If your chatbot includes a satisfaction survey at the end of conversations, track the scores over time. Compare chatbot satisfaction scores with staff scores for similar types of questions.
What to aim for: Chatbot satisfaction scores (often called CSAT scores) should be close to staff scores for the types of questions the chatbot handles. For simple factual questions, chatbot scores are sometimes higher, because the answers are immediate and consistent.
First Response Time
This is one of the most useful chatbot measures. Chatbots respond in seconds, while staff members may take minutes or hours. Track how much your average first response time improves across all customer conversations.
Lead Capture Rate
If your chatbot collects contact details, track the percentage of conversations that end with a lead collected. Compare this with the rate at which visitors complete the contact form on your website.
What to aim for: A well-designed chatbot often collects leads from a larger share of visitors than a static form does.
Calculating Total ROI
To calculate your chatbot's total ROI, use these formulas:
- Monthly cost savings = (conversations deflected per month) multiplied by (cost per human-handled ticket)
- Monthly revenue generated = (leads captured by chatbot) multiplied by (average lead-to-customer conversion rate) multiplied by (average customer value)
- Total monthly value = cost savings plus revenue generated
- Monthly ROI = (total monthly value minus chatbot platform cost) divided by chatbot platform cost, multiplied by 100
The return usually grows over the first few months, as you improve the knowledge base and the deflection rate rises.
Building a Business Case
If you need to explain the investment to owners, managers or a board, present the information in business terms:
- Current support costs and how the number of requests is changing
- Expected deflection rates, based on reasonable estimates
- Expected cost savings and how long the chatbot will take to pay for itself
- Additional revenue from collecting more leads and being available after hours
- The advantage over competitors of offering support 24/7
For many organizations, especially those receiving a growing number of support requests, the numbers favor adding a chatbot.
Start Measuring Today
Before adding a chatbot, record your starting figures: current ticket volume, cost per ticket, average response time, and the rate at which leads become customers. Then add your chatbot with Alma and measure the same figures over time to see the results.
Related reading: Chatbot Analytics: Understanding What Your Customers Really Want, Building Customer Self-Service Portals with AI.
For details, read what Alma measures out of the box.
