ROI

Measure the financial value your bot generates by resolving customer issues without human intervention

The ROI dashboard shows you how much money your chatbot helps you save by handling customer requests without needing a human.


What Counts as ROI?

Every time the bot answers a question or solves a problem without handing over to a human, it’s called a deflection — and it creates value.

But not all deflections are equal. We don’t count deflected conversations that:

  • Triggered a misunderstanding
  • Received a negative user satisfaction score

This way, your ROI reflects only the real value your bot delivers.

Customize the ROI Formula

You can set your own value for what a successful deflection is worth to your company.


How to update it:

  • Go to the ROI Dashboard.

  • Click the settings icon next to “Conversations raw value”.

  • Adjust the value for:

    • Action deflections
    • Information deflections
    • Advanced builder deflections

The default value is €4 per deflection, but you can change it anytime. Updates will apply to new conversations starting the next day.

Key Metrics

Metric

What It Means

Raw Value

The total value of all deflected conversations, before any deductions.

ROI Value

The final value after subtracting conversations with misunderstandings or negative user feedback.

Misunderstanding Rate

% of deflected conversations where the bot misunderstood the user.

Discontent Rate

% of deflected conversations that ended with a negative (or possibly negative) experience.

How ROI Is Calculated

Here’s the formula used:

ROI = Raw Value − (Raw Value × Average Error Rate)
Where:
Average Error Rate = (Misunderstanding Rate + Discontent Rate) ÷ 2

1️⃣ Raw Value Calculation:

The raw value is the total number of deflected conversations multiplied by the deflection value (default is €4 or your customized value).

Example: If your raw value is €250,988, and the total deflected conversations are 62,747, then:
Raw Value = 62,747 × 4 = €250,988*

2️⃣ Deducting Misunderstandings and Discontent:

From the raw value, subtract the value attributed to conversations that included misunderstandings or received negative user ratings. This ensures only the conversations that truly added value are considered.

3️⃣ Result

The ROI will naturally be lower than the raw value since it excludes the conversations that didn’t fully satisfy the customer.

Example: If the raw value is €250,988 and the ROI after deductions is €192,700, then:
ROI = €250,988 - (€250,988 × [(0.175 + 0.29) / 2]) = €192,700*

Example:
  • Raw Value = €250,988
  • Misunderstanding Rate = 17.5%
  • Discontent Rate = 29%
  • ROI = €250,988 − (€250,988 × (0.175 + 0.29) ÷ 2)
    Final ROI = €192,700

How Misunderstanding & Discontent Are Measured

  • Misunderstanding Rate
    % of deflected conversations where the bot misunderstood the request.
    Formula: misunderstood deflections ÷ total deflected conversations
  • Discontent Rate
    % of deflected conversations that left users unhappy.
    Formula: (Negative CSAT + 20% of Unrated CSAT) ÷ all CSAT responses

🤔 Why count 20% of unrated CSAT as negative?

Because people often skip feedback when they’re unhappy. This adds realism to your ROI score.


Next page