The reference rate is a process indicator, not a number for which the boss will approve the budget. What really needs to be in the center of the dashboard is "this month AI quotes the potential business opportunities for us to bring NT$." Give a single percentage of the reference rate to the supervisor, mostly by saying, "So?" It gives you a set of fields and equations that can be reproduced directly, so that visibility can be changed from an abstract ratio to a line amount that the finance minister can understand.
Why does "citation rate" mean nothing alone?
The AI engine changed the user's journey. They used to search for results on your website, and now a lot of people end up asking questions in ChatGPT, Perplexity or Google AI Overviews, without any connection. This represents traditional natural flows, keyword rankings, which underestimate your real visibility. The reference rate on this is measured by whether AI generates an answer that takes you as a source and mentions your brand. The problem is that 18% of this number has no weight for decision makers. It doesn't tell the executive whether to put another season budget or not, and it can't talk to the line of business. In order for the dashboard to be taken seriously, you have to put the reference rate in a chain to the harvest.
Six fields to track on the dashboard.
Don't do 20 sign panels. No one will look. One table, six columns, each theme group in a row, is enough to sustain the weekly decision.
- Target theme/Query group: grouping questions that buyers really ask, such as "B2B SaaS Import Costs" instead of a single keyword.
- AI Reference Rate (Strategic Platform): ChatGPT, Perplexity, Google AI Overviews, Gemini each write a column because their origins are much different.
- Position of brand in the reference: listed as the first chosen source, only mentioned or not at all, with three values far apart.
- The correctness and emotionality of the reference: AI's statement about whether your content is correct, positive or negative, and the incorrect reference is uniquely marked.
- Monthly Business (NT$): This column is for the supervisor, the formula is in the next paragraph.
- Week-to-week change and gap theme: Which issues fell out and which new entries were added, deciding on the next week's schedule.
The first four columns are original observations, the fifth column is the result, and the sixth column is the basis for action. With the fifth column missing, the dashboard was just a good surveillance report; with the sixth column missing, there was no idea what to do next.
The formula for converting references to NT$
The core formula is only one, written into a trial scale, and each theme group is available once: monthly AI business (NT$) = target theme monthly query x AI reference x reference exposure rate x interview lead rate x lead business rate x average business volume. The preceding paragraph is visible data (characteristics, reference rates), the middle paragraph is behaviour conversion (pointing, leads, business), and the last paragraph is a financial path (average cost per user unit). Each one of them is capable of being verified or used as a conservative representation, and cannot be filled in by feeling. The real value is not the absolute value of a single month, but when you recalculate it with the same calibration every month, the trend becomes evidence that you can talk about the budget.
A reality test.
Walk through a set of illustrative numbers, you'll know the weight of each column. The theme group "B2B SaaS Import Costs" is estimated at 8,000 per month; your AI reference rate is 18 per cent, with about 1,440 references; the reference exposure rate is 12 per cent, with about 173 interviews; the call lead rate is 4 per cent, with about 7 leads; the call switcher rate is 30 per cent, with about 2 businesses; the average business balance is $180,000. After multiplying, this theme group is given a single month to calculate a potential pipeline of approximately $370,000 NT. If you add ten theme groups to the list, there's a number at the bottom of the dashboard that the supervisor is willing to spend time discussing. And more importantly, when the reference rate falls from 18% to 12%, a third of the business opportunities in this column are missing, the supervisor can see the loss at once, without you explaining it.
Update the rhythm and attribution.
The dashboard will expire, usually not by mistake, but by no means, or by no means by any means. It's like writing the rhythm and the person in charge at the top of the document.
- Monday: The content manager pulls raw data on the citation rates of each platform to mark new and falling themes.
- Weekly: Marketing converts the theme of the gap to next week's content or fixes the schedule so that observations do not turn.
- Monthly: Recalculate NT$ business opportunities on all topics, and match the actual trade with the industry system, return conversion rate scenarios.
- Type alignment: Average commercial amounts and commercial definitions are industry-specific, avoiding marketing and accounting for business.
Monthly return to school is the easiest to skip and the most important. The first version assumes that it will not be possible, but once the real deal is fixed every month, it'll be clearly closed in three or four months, and the board's persuasive force will grow.
Three most common mistakes
First, the reference rates of the four platforms are mixed into an average. ChatGPT, unlike the source system of Perplexity, can't figure out which platform to fix first. Secondly, only the reference rate is tracked and the correctness of the reference is not tracked. AI quoted you, and made a mistake about the price or the function, which was a negative asset, which had to be singled out. Thirdly, the dashboard and the line of business are separate and separate, and the manager is unable to judge the cause and effect of the sale of another set of transactions with a visible number and a newspaper. These three points point to the same thing: if the visibility data are not even on the way to the airport, the fine graphics cannot afford a budget conference.
If you can't see it, you can't be converted into money, you'll always be just a sales team that feels good.
You'll start with these six columns and one formula to put up the dashboard for two months, and you'll soon know which subjects you're using as a source and which high-value topics you're missing. If you want to clear the gap faster and see how far your brand is being quoted in the various AI engines, you can expect a 30-minute GEO diagnosis, and we can match your first version with Brand Radar's actual data.
