Guide
What You Can Export, and Where
Everything indextr measures can leave with you: one PDF report and four spreadsheets. This page says where each button is, what every column holds, and what a blank cell is telling you.
The short version
Exports are on every paid plan. The PDF is the page you are looking at, printed clean, for a client or a web developer. The spreadsheets are the same measurements as rows and columns, for sorting, filtering and keeping alongside your other reporting. Each file is named with your company, what it holds and the date it was made.
One rule runs through all of them: a file always agrees with the screen it came from. The same answers are counted, the same ones are left out, and the same words are used. If the screen says Prominence, so does the file.
Where the buttons are
All of them sit in the top-right corner of their page, beside the page title.
Sidebar, then Analysis
- Print / Save PDF
- Download score history (one row per scan)
- Download by engine (one row per scan per engine)
Sidebar, then Scan History, then open any scan
- Download full results (one row per phrase per engine)
- Print / Save PDF
Sidebar, then Recommendations
- Download CSV
On a narrow window the buttons drop under the page title rather than staying to the right of it. On the free plan they are not shown, because the free plan is diagnosis only.
Full results of one scan
The raw measurement. One row for every search phrase on every engine the scan ran, sorted by phrase and then engine, the way the scan report reads. This is the file for anyone who wants to check the work, and the one that holds the most.
The phrase that was asked, exactly as you track it.
ChatGPT, Google AI, Gemini, Claude or Perplexity.
yes if the answer named your business, no if it answered and did not.
yes if the answer linked to your website as a source.
Your place when the answer was a list: 1 is first. Blank when the answer was prose or did not name you.
positive, neutral or negative, read from the wording of the answer itself, only where you were named.
yes, or not measured when the engine gave no answer to read, for example asking for your location instead. Those rows are left blank rather than counted as a miss.
Every business the answer named, separated by semicolons. This is who the engine recommends on that phrase.
Every web address the engine drew on to answer, separated by semicolons: your site, directories, competitors, trade press. This is what it actually read. It says none when the answer cited nothing, and is blank for scans made before sources were kept.
One column per competitor you track: yes if that answer named them, no if it did not, blank if the scan never looked for them. When you add a competitor, past answers are re-read for that name, so the column fills in for earlier scans.
Branded phrases, the ones that name your business in the question, are in this file and marked on the Searches page. They are measured but never scored, so they are not in the by-engine file below and do not move your Visibility Score.
Score history
One row per scan, oldest first: the trend chart as a table. A scan you have excluded from the trend is left out here too, so the file and the chart never disagree.
The day the scan was scored.
The headline number, 0 to 100.
The share of all answers in that scan that named you.
Of the answers that named you on engines scored for citations, the share that linked your site.
0 to 100 across the answers that named you: 100 is every one positive, 50 is neutral.
How often you were placed ahead of the competitors named in the same answer. Shown for context; it does not move the score. Computed when each scan ran, against the competitors tracked at the time, so a competitor added later does not change earlier rows.
The largest part of the score: 100 for first place, 80 for second or third, 60 for fourth or lower, whether the answer was a list or prose named in that order; 70 for a mention whose order could not be read; and 0 for a search that did not name you. Averaged over every scored search.
One column per engine, the share of that engine's answers that named you. Blank when the engine did not run in that scan, which is what a plan without it looks like.
By engine, over time
The By Engine table from the Analysis page, for every scan rather than only the latest. One row per scan per engine, so you can filter to one engine and see whether ChatGPT is improving while Gemini stands still. It is the file for the question “where am I actually losing?”
The day the scan was scored.
One row for each engine that ran in that scan. An engine that did not run has no row rather than a row of zeros.
How many answers that engine gave that could be measured, branded phrases left out.
The share of those answers that named you.
Same scale as in score history, for this engine alone.
Of the answers that named you, the share that linked your site. Blank for Gemini and Claude, which the score does not count citations for. Gemini's citations are mostly YouTube and Reddit rather than the business named, which is why it was left out; Claude was never included.
0 to 100 across the answers that named you. Blank when none did, because there is no wording to read.
These rows do not add up to the Visibility Score, and are not meant to. The score is a weighted average, with ChatGPT and Google AI carrying most of it, so the same gain on two engines moves it by different amounts. Read this file for which engine moved, and the score history for what that did to the number.
Recommendations
Your action list, with the steps, in a form you can hand to whoever does the work or track in your own task list. Retired recommendations are left out.
new, completed or dismissed.
1 is the one to do first.
The recommendation and the paragraph beneath it.
The implementation steps, one per line inside the cell. Spreadsheets show the cell as several lines.
content, technical, reviews, directory, citations or qa.
quick, medium or long: when to check back, not when something will move. No engine publishes how often it re-reads a page.
What the recommendation is meant to move: mentions, citations, sentiment or position. Blank where we would only be guessing.
Which engines it is aimed at, or all engines.
When the recommendation was written, and when you marked it done.
What a blank cell means
Never zero. Every file follows the same rule the dashboard does: a number is written only where something was measured. An engine that gave no answer to read leaves its row blank and says not measured. An engine your plan does not include leaves its column blank or has no row. Sentiment is blank where you were not named, because there is no wording to judge.
Zero is written in one case only: the engine answered, and did not name you. That is a real finding, and the one worth acting on.
Grey on screen, counts in the file
On the dashboard and the Analysis page, a rate resting on fewer than 6answers is shown in grey, with no red, amber or green and no change arrow. One answer flipping would move such a rate by 20 points or more, so a colour on it would be a colour on a single answer. The Visibility Score’s change is treated the same way: a move smaller than what one answer could cause is marked “within one answer” rather than shown as a rise or a fall.
A spreadsheet has no grey. What it has is the count: the Answerscolumn in the by-engine file, and the “n of N” under every rate on the pages. Read a rate together with its count. Two of three answers is one answer’s opinion; twenty of thirty is a pattern.
The 3-scan average drawn on the trend chart is not a column in any file, because it is arithmetic on columns that are: average each scan’s Visibility Score with the two before it in the score history file and you have the same line.
The PDF
Print / Save PDF opens your browser’s print dialog with the dashboard navigation and buttons removed. Choose Save as PDF as the destination. It is the same page you are looking at, so it holds whatever the page holds on the day: the Analysis page gives you the trend, the engines and the phrase table, and a scan report gives you every answer from that scan with the recommendations it produced.