augenmerk Documentation / Findings catalogue

9 · Documentation

Findings catalogue

All findings augenmerk derives from a run, with reason, action and the number that shows the effect.

A finding is a rule over the measurements, not a list someone maintains. After every run the application checks the whole catalogue against the project's numbers. Whatever applies appears under "Next steps", ordered by effect.

Order

Cause before gap before evidence before channel before position before movement before source before topics. Within the same type the larger deviation counts. A cause, such as a blocked AI bot, always comes first: as long as it persists, everything else is secondary.

TypeWhat it describes
CauseSomething prevents the AI from reading the website.
GapThe brand is missing somewhere it could stand.
EvidenceThe statements about the brand have too few or only own sources.
ChannelThe channels see the brand differently.
PositionWhere the brand stands in the ranking and how securely.
MovementWhat has moved above the noise band since the last run.
SourceWhich domains shape the market.
TopicsWhat the AI praises, criticises or links to others.
ResolvedA cause that has disappeared since the last run.

The file

Every finding has a file with four sections: the finding as a sentence with the real numbers, why it matters, what to do, and how to see the effect. The last section names the number that has to move, where it stands and when a change can be expected. Resolved is determined by the measurement, not by the user: as soon as the number moves, the finding goes to "Resolved" and stays visible for one run. Findings without an action are observations; their file says what would tip the finding.

For competitors viewed through the lens, the same observations exist, but no actions.

The catalogue

In the sentences, placeholders in angle brackets stand where the application shows the project's numbers and names.

Cause

Your website rejects automated requests (status status).

Why it matters
A page the AI is not allowed to read cannot be used as a source either. Everything the search channels say about you therefore comes from other sites.
What to do
Open AI access. Check with your IT whether bot protection or a firewall is rejecting the AI bots. The website view shows which bot gets which response.
How you will see the effect
The fetch status in the website view, today status. As soon as a run sees status 200, the finding moves to “Resolved”. Mentions of your website in search answers follow two to three runs later, first on Perplexity and Google.

gesperrt of gesamt AI bots are blocked in your robots.txt.

Why it matters
A blocked bot does not read the page and cannot cite it, even if Google has known it for years.
What to do
Review the blocks. Decide per bot whether the block is intended. The list is in the website view.
How you will see the effect
The number of bots allowed to read, in the website view, today gesamt minus gesperrt. Every opened block counts in the next run; citations from the affected systems follow two to three runs later.

bots AI bots may read only parts of your website.

Why it matters
Blocked paths include pfade. If product or help pages sit there, AI lacks them as sources.
What to do
Review blocked paths. Open what should serve as a source and keep closed what should not leave the house.
How you will see the effect
The blocked paths per bot in the website view, today bots bots with restrictions. Opened paths appear as sources from the run after next, if the pages there have something to say.

Gap

Mentioned in none of n answers to segment questions.

Why it matters
Instead, the AI names marken. Usually it is one of two things: your website does not describe the offer clearly, or the sources the AI reads from do not list you in this segment.
What to do
Check segment presence. First check whether the offer is clearly named on your website, then which sources AI uses for these questions.
How you will see the effect
Presence for segment questions in the questions view, today 0 of n answers. A single mention may be chance; from two runs with mentions the gap counts as closed.

Only x % for segment, that is y percentage points below the brand value.

Why it matters
The brand is known, but hardly in this segment. Questions about it are answered with other names.
What to do
Check segment pages. See whether the product line has its own clearly named page and whether AI's sources list it.
How you will see the effect
Presence for segment in the questions view, today x %. The finding is resolved when the gap to the brand value drops below the noise band, not after one good run.

Range presence is y percentage points below the brand value: sortiment % compared with marke %.

Why it matters
The AI knows the brand, but not all its product lines equally well. The good brand value hides that.
What to do
Look at product lines one by one. The “Presence per product line” card shows which lines carry and which are missing. Start with the line with the largest gap.
How you will see the effect
Range presence on the overview, today sortiment % versus marke %. The finding closes when the gap stays below the noise band for two runs.

For “frage” the brand is not mentioned on any channel, not in laeufe runs.

Why it matters
A counting question without a single mention is no coincidence but a gap in what AI knows or finds about you.
What to do
Answer this question on your own website. Check which page of your website answers this question — and which sources AI cites instead.
How you will see the effect
This question's row in the questions view, today laeufe runs without a mention. One mention in one channel makes the finding disappear; it is reliable only after two runs.

Your page for “suchbegriff” ranks platz on Google, but is cited by no AI.

Why it matters
suchen searches a month and not a single AI citation: the demand is there, but the AI does not use the page as a source. Path pfad.
What to do
Make the page readable for AI. Put the answer to the search question at the top, clear structure, references, and check whether AI bots are allowed to read the page.
How you will see the effect
AI citations of this page in the gap view, today 0 with suchen searches a month. Expect the first citations two to three runs after the change, first on Perplexity and Google.

n of your m strongest Google pages are cited by no AI.

Why it matters
A good Google ranking does not yet mean the AI uses the page as a source. Exactly this gap is the most concrete to-do list there is.
What to do
Go through the strongest pages. The website view puts rankings and AI citations side by side per page. Start with the page with the most searches.
How you will see the effect
The number of uncited pages in the gap view, today n of m. Every page that gains a citation leaves the list in the following run.

Your home page contains n internal links. An AI reading the page mostly sees navigation.

Why it matters
AI systems read a page top to bottom and often only the first few thousand characters. With more than 60 links, about a third of that is left for the actual content.
What to do
Slim down the navigation. Put content before link lists, shorten or lazy-load mega menus and footer directories. We recount the links after every run.
How you will see the effect
The number of internal links on the home page in the website view, today n. We recount after every run; below 60 the finding is resolved.

The website is losing on Google: keywords % fewer keywords and besuche % fewer visits in monate months.

Why it matters
The search channels read the pages that rank up front on Google. If the website loses there, it also loses, with a delay, as a source for the AI. Whether season, relaunch or lost rankings are behind it, the history does not say.
What to do
Check the Google history. Check in Search Console when and on which pages the decline began. The website view shows keywords and visits per month; the gap view shows which pages the AI still cites.
How you will see the effect
Keywords and visits per month in the website view, today keywordsBis keywords and besucheBis visits. Google data arrives monthly, not per run; the finding disappears when the decline over the shown period drops below 30 %.

Your website is cited in x% of search answers, domain in y%.

Why it matters
The picture the AI paints of you is formed on other sites, not on your own.
What to do
Look at domain. See what domain writes about you and your competitors, and whether you are missing there.
How you will see the effect
The share of search answers citing your website, sources view, today x %. Movement shows run by run; the finding is resolved when domain no longer leads you by more than the noise band.

Evidence

Everything AI says about you with a source comes from domain.

Why it matters
belegt statements with a source, not a single one from an outside source. The AI backs your competitors from portale; for you, such independent confirmation is still missing.
What to do
Build independent sources. Check whether the portals AI uses to back your competitors list you currently and completely.
How you will see the effect
Statements with an external source in the topics view, today 0 of belegt. The first statement the AI backs about you from a portal resolves the finding.

Only x% of statements about you have a source.

Why it matters
belegt of gesamt. What the AI says without a source comes from model knowledge and can change with the next model version without you having done anything.
What to do
See statements with a source. The topics view separates statements with and without a source. See which sites the few sources are.
How you will see the effect
The share of statements with a source in the topics view, today x %. It rises when the search channels find pages that write about you; a movement above the noise band counts.

kanal names a source for x % of its statements about you, kanal2 only for y %.

Why it matters
Same market, two systems, two ways of working: where few sources are named, model knowledge is speaking.
What to do
Compare sources per channel. The share with source per channel in the topics view shows which system reads from sources and which answers from model knowledge.
How you will see the effect
The share with a source per channel in the topics view: kanal x %, kanal2 y %. This is how the systems work, not a task; the finding stays as long as the spread exists.

Channel

Known in model knowledge (g %), mentioned less often in answers with live search (l %).

Why it matters
The AI knows you, but does not find you in the sources it currently reads. That is a source problem, not an awareness problem.
What to do
Check sources behind buying advice. Which test and comparison sites do the search channels use, and which brands do they name instead of yours?
How you will see the effect
Presence in live search on the overview, today l % versus g % in model knowledge. Resolved when the gap drops below the noise band.

Mentioned more often in live search than in model knowledge: l % compared with g %.

Why it matters
The sources know you, the models not yet. Typical for young brands; model knowledge catches up with the next model versions.
How you will see change
Presence in model knowledge, today g %. It catches up with new model versions, not with runs; think in months.

engine mentions you in x % of answers, the other systems in y %.

Why it matters
A single system almost overlooks you. Whoever asks there hears other names.
What to do
Read this channel's answers. The answers view shows the wording per channel — see which brands engine names instead.
How you will see the effect
The presence of engine in the answers view, today x %. A single run fluctuates; the finding is resolved when engine sits closer to the other systems for two runs.

engine mentions you in x % of answers, the best search channel in y %.

Why it matters
Search channels read different sources. This one apparently reads pages where you are missing.
What to do
Check this channel's sources. The sources view shows per channel which domains are cited.
How you will see the effect
The presence of engine, today x % versus y % in the best search channel. Movement follows the sources this channel reads; the sources view shows them per channel.

x % in Google's answer box, y % in the live searches of the chat systems.

Why it matters
Google draws its answer from its own search, the chat systems from their sources. Two views of the same market.
What to do
Compare Google rankings with AI sources. The website view shows which of your pages rank on Google and whether AI cites them.
How you will see the effect
Presence in Google's answer box, today x %, against the live searches at y %. Google follows its own rankings; movement shows where your pages rise or fall on Google.

Position

Not among the marken most-mentioned brands of the market.

Why it matters
A brand enters the ranking from two mentions in counting answers. Below that, one cannot yet speak of presence.
What to do
Go through the questions. See in the questions view where the brand appears at all.
How you will see the effect
The ranking on the competition page, today without your brand. From two mentions in counting answers you appear there; the finding then turns into a position statement.

Rank rang, but the lead over marke is only d percentage points, less than the fluctuation range of rauschen.

Why it matters
In the next run the order can flip without anything really having changed.
What to do
Keep watching. Only a movement beyond the fluctuation range is a real movement. The competition view shows the course of both brands.
How you will see the effect
The gap to marke in the competition view, today d percentage points with a noise band of rauschen. Only a gap above it, for two runs, is a secure place.

Most mentioned of marken brands, d percentage points ahead of marke.

Why it matters
The lead is beyond the fluctuation range; it is reliable.
How you will see change
The lead over marke, today d percentage points. If it drops below the noise band, the overview reports a wobbling place.

Rank rang of marken. d percentage points short of rank vor (marke).

Why it matters
The next rank is within reach. The competition view shows on which channels marke gains its lead.
What to do
Look at the gap per channel. Where marke leads and you do not, the lever sits.
How you will see the effect
The gap to marke in the competition view, today d percentage points. If it closes by more than the noise band, place vor is in sight; a change of place appears as movement on the overview.

Rank rang of marken. d percentage points short of rank vor (marke).

Why it matters
A large gap rarely arises in one place, but across many sources. The competition view shows where.
What to do
See competition. Compare channels and product lines; the gap is never the same size everywhere.
How you will see the effect
The gap to marke, today d percentage points. It closes across sources and product lines, run by run; the competition view tracks both brands.

When mentioned, on average at rank platz, calculated from n placements in ranking questions.

Why it matters
AI knows the brand but lists it far down. Whoever reads only the first three does not see you.
What to do
Check positioning. The competition view shows which brands the AI names ahead of you; the answers show how it justifies that.
How you will see the effect
The average place in ranking questions, competition view, today platz from n placements. A value below 3 puts you in the part of the list everyone reads.

Movement

Presence up d percentage points since the last run.

Why it matters
The movement is beyond the fluctuation range, so it is reliable.
What to do
Find the cause. The questions view shows for which questions the brand is newly mentioned.
How you will see the effect
Whether the gain holds: presence in the next run. If it keeps the new value, the movement is real; if it falls back, it was an outlier.

Presence down d percentage points since the last run.

Why it matters
The movement is beyond the fluctuation range. Something has changed in the answers or in the sources.
What to do
Find the cause. The questions view shows for which questions the brand is no longer mentioned.
How you will see the effect
Whether the drop holds: presence in the next run. A second run below the old value confirms the movement; the questions view shows where the mentions are missing.

From rank von to rank nach.

Why it matters
A change of more than one position since the last run.
What to do
See competition. Which brand swapped places, and in which channel?
How you will see the effect
The ranking in the next run. If place nach holds for two runs, the change is reliable; the gap to the neighbour is in the competition view.

marke has gained d percentage points since the last run.

Why it matters
A reliable movement for a competitor is a signal: somewhere a source or a campaign is working.
What to do
See marke's sources. Select marke as the brand at the top; the sources view then shows where the gain comes from.
How you will see the effect
The presence of marke in the next run. If the gain stays, a new source has taken hold; the sources view in the marke lens shows it.

No movement beyond the fluctuation range for n runs; presence stable at sicht %.

Why it matters
Steady means: what you see is reliable. Changes from here are real changes.
How you will see change
Presence on the overview, today sicht %. From now on every movement above the noise band appears as its own finding.

Source

domain is, with n answers, the most-cited source in your market. Your website is not among the first ten.

Why it matters
Whoever writes the buying advice shapes the answer. This source shapes your market.
What to do
Check presence on domain. In which articles do competitors appear, in which do you? The sources view shows every answer.
How you will see the effect
The occurrences of domain in the sources view: answers in which you are named alongside the competitors. Today n answers with this source. A place among your website's top ten sources resolves the finding.

domain is cited when the AI names wettbewerber, for you not a single time in n answers.

Why it matters
A source that knows the market, but not you. This gap can be closed.
What to do
Search domain for the brand. If your brand is missing there, this site is the first contact for getting into tests or comparisons.
How you will see the effect
The occurrences of domain in the sources view, filtered by “named”, today 0 of n. As soon as domain names you in one answer, the finding disappears.

From “frage” AI makes the search teilfrage.

Why it matters
Before the AI answers, it searches the web with these words. Whoever ranks high on Google for this search is read as a source.
What to do
Check the ranking for this search query. Check whether one of your pages appears on Google for “teilfrage”. The website view shows your rankings.
How you will see the effect
Your Google rank for “teilfrage” in the website view and the sources of the answers to “frage”. If one of your pages ranks up front for this search, it should appear as a source one to two runs later.

Your website is cited in x % of search answers, n of live.

Why it matters
The AI uses your pages as a source. Many brands are still working towards that.
How you will see change
The share remains the yardstick: today x %. If it falls by more than the noise band, the overview reports the movement.

Topics

AI praises above all thema: zitat

Why it matters
The passage comes from an original answer. This is how the AI describes your strength, in its own words.
How you will see change
The statements on thema in the topics view: their number and their share of praise. If both hold, the topic carries; if it tips, the overview reports the criticism.

Criticism above all for thema: zitat

Why it matters
The passage comes from an original answer. The topics view shows whether the criticism is sourced and from where.
What to do
Read the criticism. Criticism with a source can be addressed at the source; criticism without a source comes from model knowledge and changes only slowly. The topics view separates the two.
How you will see the effect
The statements on thema in the topics view, filtered by criticism. Criticism with a source disappears when the source changes; criticism from model knowledge takes months.

thema is praise and criticism at once: praised lob times, criticised kritik times.

Why it matters
AI is not of one mind. Usually it depends on which source it is reading.
What to do
Check sources on the topic. Which sources carry the praise, which the criticism? The topics view shows each statement with evidence.
How you will see the effect
The ratio for thema in the topics view, today lob praise to kritik criticism. Resolved when one side clearly prevails for two runs.

x % of statements about you are negative, n of gesamt.

Why it matters
That is more than usual: in AI answers about four in five statements about brands are positive.
What to do
Read the negative statements. The topics view shows every negative statement with its passage and, where present, its source.
How you will see the effect
The share of negative statements in the topics view, today x %. Below about 20 % it is within the usual; the finding then disappears.

On thema you are the only brand that is criticised: x % of your n statements on it are criticism, while marken competitors are praised.

Why it matters
A topic the AI frames positively for everyone else and negatively for you. That shapes the comparison more than any metric.
What to do
Read the topic in competition. The topics view shows every statement with passage and source, the competitors' too. What do their sources say on the topic, and what is missing from yours?
How you will see the effect
The matrix in the topics view for thema: your cell against the competitors'. Resolved when your statements on it are mostly praise or the topic is no longer negative for you alone.

On thema you are the only brand the AI praises: x % of your n statements on it are praise, marken competitors are criticised.

Why it matters
A topic where the comparison clearly goes your way. Such leads are rare and carry the recommendation.
How you will see change
The matrix for thema: if your cell stays praise and the competitors' stays criticism, the lead holds. New statements arrive with every run.

On thema you are the most-mentioned brand: n statements, x % of them praise. marke follows with m.

Why it matters
The AI links the topic mainly with you. That is the kind of visibility that decides recommendations.
How you will see change
The number of statements on thema per brand, today n against m for marke. If the gap closes, the overview reports the topic as contested.

On thema the AI talks about marke (m statements), hardly about you (n).

Why it matters
A topic that matters in the market but that the AI does not connect with you. Missing here means missing from the answers to that need.
What to do
Claim the topic. The topics view shows which sources carry marke on this topic. A page of your own that answers the topic clearly and verifiably is the first step.
How you will see the effect
The statements on thema about you in the topics view, today n against m for marke. First statements show once the search channels read a page on it, usually two runs after publication.

On thema your brand is praised (y %), marke criticised: x % of m statements.

Why it matters
A lead over a competitor that the AI names on its own. Contrasts like this can be widened.
What to do
Widen the lead. Back the topic more strongly on your own website with comparisons, tests and figures. The topics view shows which sources carry the criticism of marke.
How you will see the effect
The matrix for thema: your praise (y %) against the criticism of marke (x %). If the contrast holds, it carries the recommendation; if it tips, the overview reports it.

Resolved

Your website is reachable for automated requests again (status status).

Why it matters
In the previous run it rejected requests. From now on the search channels can use it as a source; the effect shows in the next runs.
How you will see change
Mentions of your website in search answers, sources view, row 4. We expect the first ones within two to three runs; no effect until then is normal.

erlaubt of gesamt AI bots may read now — vorher before.

Why it matters
The robots.txt has been opened since the previous run.
How you will see change
Citations from the systems whose bots may now read (erlaubt of gesamt). Search channels re-read a page only on their next visit; allow two to three runs.

Updated: 2026-09-16