◇Visibility

You type in a web address and within a minute you have Findings: what is amiss technically, in terms of content and legally — with instructions for each point.

Three areas are checked: business data for search engines (contact details, local relevance, preview when sharing), the Content, and the legal side — legal notice, privacy notice and services that send data to third parties without asking.

The check becomes a project, and that is where it continues. The tabs are called “Findings”, “Competition”, “Keywords”, “Market radar”, “AI visibility” and “Backlinks”. Under “Keywords” you find rank measurement, search volume and the visibility index.

Type in the address and town, then check. Aurora Hive Sales reads the pages itself — that costs nothing and may run at any time.

Depending on the other server, reading takes up to a minute and a half. The first check automatically creates an entry that you can find again later.

Enlarge

The check itself does not. Whatever costs money is switched off by default and is listed with a price tag in the settings of the area:

• Rank measurement, that is the ranks for your keywords, plus the entry in the map service and the Market radar. They use research queries, because they query the paid Google search.
• The weekly measuring run, which repeats a project’s rank measurement by itself.
• The Backlinks tab. Every retrieval uses data queries.
• AI visibility, whose measuring runs have their own budget.

There is one exception without a switch: you fetch the Search volume with its own button, and in front of the button it says what it costs. So nobody slips into it by accident.

Each point carries three pieces of information: what has been found, why it matters, what to do.

The third part is the reason the whole thing is useful: findings without instructions would just be a complaint.

And if there is nothing to do, it says so too — “There is no point that needs to be tackled immediately.”

Yes, as a PDF report — the same result, but in the language of a master craftsman rather than a technician. At the end it says what is already working.

This report is the item you hand over: to a customer, to an agency, to your own management.

Whatever has already been measured in the project goes into it. Observed keywords appear with their rank and, once determined, with the searches per month. If ranks have been measured, the visibility index is included, with a small trend over the last few weeks and the strongest competitors. And if a backlink retrieval exists, a short section on the links from other websites follows.

What has been checked once stays. With the second check you see what has changed.

That is the proof after a rebuild — and, if you sell this work, the best argument for the next invoice: not the claim that things have got better, but the side-by-side comparison.

In this order, and it saves money:

1. Get suggestions. They come from Google’s own autocomplete and cost nothing. How often a term appears there gives a first, free indication.
2. Determine the search volume if needed. Below the suggestions there is “Determine search volume”: Google’s estimate of how often each term is searched for per month. The whole list costs one data query; afterwards the recommendation and topics are ordered by volume.
3. Adopt the suitable ones and only then measure. Measuring the ranks costs one research query per keyword.

Hence the rule: few, and the right ones. Measuring twenty keywords of which fifteen nobody searches for is the most common mistake. With the search volume you see such terms before they cost money.

A note in the display always states how many queries measuring would cost at that moment.

Not a picture of the whole web, but a complete index of your segment: who appears in your region for your keywords, how good their sites are, and which directories everyone is listed in except you.

The last one is usually the most useful: a directory in which all your competitors are listed and you are not is a task for the next morning.

Prerequisite: create keywords first, then build the radar. Without keywords the radar has nothing to search for.

Yes — enter up to three and then start a check; they are then checked as well.

If the radar says your address does not appear in the queried search results, there are two possible reasons: either you are present everywhere, or too few competitors have been read so far. A look at the checked keywords tells you which.

Two common causes:

• The address is wrong. Check it in the browser. Some websites have no reachable home page at the address you would expect.
• The site requires a sign-in. Pages that can only be reached when signed in cannot be assessed.

For the speed measurement there is a third cause: it comes from an external measuring service, and without your own access, all queries worldwide share a common daily quota — which is practically always used up. Your own access for it is free of charge and is stored once; ask your contact person. Everything else in the check works without it too.

Yes — only what is publicly accessible anyway is checked; Aurora Hive Sales reads the pages like a visitor.

Two things still apply: findings about someone else’s site are a statement about its operator — treat them fairly if you use them as a sales argument. And you should not run the paid parts (rank measurement, market radar) on spec for twenty competitors.

As an item, not a lecture: the PDF report is what you hand over — in the language of a master craftsman, and with what is already working at the end.

Three points that help:

• Legal findings first. A missing legal notice is a concrete risk, not a matter of taste.
• Not everything at once. Three points that can be done this week have more effect than a list of thirty.
• Second check after the rebuild. The comparison is the best argument for the next invoice.

Legal notice, privacy notice and services that send data to third parties without asking.

The last item in the list is the one that comes up most often and is least known: fonts, maps, videos and icon sets from third-party servers transmit every visitor’s IP address there — open to challenge without consent.

This can almost always be fixed by delivering the same files yourself.

What search engines need in order to understand a page: business data for search engines, contact details, local relevance and the preview when sharing.

The last point is easily overlooked and has an immediate effect: when someone shares your address in a messenger, the stored preview decides whether an image with a title appears there or a bare address.

The speed measurement is a special case — it needs its own free key; everything else works without it.

The free check may run at any time — it makes sense after every rebuild and once a quarter.

With the paid parts, restraint pays off. For most businesses one rank measurement a month is enough to see movement. Every measurement costs one research query per keyword, so use few keywords, and the right ones.

Anyone who wants to follow the visibility index as a weekly curve can switch on the weekly measuring run for a project. Individual swings from one week to the next are often chance; the direction over several weeks is what tells you something.

That is shown by the Search volume: Google’s estimate of how often a term is searched for per month, averaged over the last twelve months. It appears as “approx. 1,900” because Google itself rounds to bands. Especially with small values, “approx. 70” means roughly seventy, not exactly seventy. Aurora Hive Sales does not round any further and does not extrapolate anything.

It is fetched in two places, each at the press of a button:

• Below the keyword suggestions with “Determine search volume”. One data query covers the whole list of suggestions.
• For the observed keywords with “Update search volume”. The value then appears in the “Searches/month” column. If keywords have been created for several towns, each town costs one data query.

Without this click, you only have the frequency from Google’s autocomplete. It costs nothing and is often enough to weigh terms against each other.

Not necessarily. “no value” means that Google gives no figure for this term, and that is something different from zero.

Two reasons are common. Google leaves very rare terms empty. And Google groups very similar spellings together, with only one of them carrying a value.

For a trade business, a term without a value can still be the right one, for example “emergency heating repair” with your own town name. Anyone searching like that usually wants to hire someone straight away.

If, on the other hand, it says “Not determined yet.”, no search volume has simply been queried for the term yet.

A list costs one data query, whether it contains three or three hundred terms. Next to the button, before you click, it says how many data queries would be due and how many you have left this month.

Values once determined are stored for 30 days. If you query the same term for the same area again within this time, even from another project, the value comes from storage. Then it says “free of charge”. After the 30 days, it is queried afresh.

If the button is greyed out, either the monthly budget is no longer sufficient, or no access to the data service has been set up for your instance. The line below the button says which of the two cases applies.

The Backlinks tab in the project answers that. With “Retrieve profile” you get:

• the number of referring websites (every other website counts once, however often it links) and how many of them link with full value, that is without a nofollow marker,
• the backlinks in total,
• a Strength from 0 to 100 that rates how well the website itself is linked,
• newly added and lost referring websites of the last 30 days,
• the strongest referring websites with anchor text, target page and the day the link had first been seen.

Two further buttons go beyond your own profile: “Compare with competitors” puts your figures next to those of up to ten competitors, and “Find gaps” finds websites that link to your competitors but not to you.

The tab is switched off by default, because every retrieval costs data queries.

It is switched on in the settings of the Visibility area via “Backlinks (who links to the website)”. Only an administration account may do this.

Every button shows its price tag beforehand:

• Retrieve profile: 3 data queries
• Compare with competitors: 3 data queries, whether two or ten competitors
• Find gaps: 1 data query

Every retrieval is saved as a snapshot. The next time you open it, you see the latest snapshot without paying again, and the history is built from the saved profiles. Nothing is repeated by itself: a new snapshot is only created when someone clicks.

If the tab reports “No data access stored.”, your instance lacks access to the data service. Saved snapshots remain visible all the same; only your provider can set up the access.

With “Find gaps” in the Backlinks tab. It looks for websites that link to at least two of the chosen competitors and not to you. Each gap is shown as a sentence, for example: seven of your competitors are linked from a business directory, you are not.

You choose the competitors at the top of the tab. The businesses entered in the project, those from the market radar and those from your rank measurements are suggested free of charge. The list deliberately does not suggest portals and directories, because they are not competitors. You can add or remove freely; on your next visit, the selection you used last is shown.

If the same address also appears among the listing gaps of the market radar, the sentence says so. Then two independent sources confirm the same gap.

Because the figures come from a backlink index that regularly reads in large parts of the web, but not the whole web. Small, young or rarely linked sites are sometimes missing from it, and this happens especially with trade businesses.

For Comparison the figures are still well suited: all competitors are measured from the same data. The trend over several retrievals is also meaningful. You should not read them as the absolute truth about every single link.

Two further points of context:

• The Strength (0 to 100) is a rating by the index and not a figure from Google.
• A link with a nofollow marker counts towards the referring websites, but is not considered a full recommendation.

A number from 0 to 100 for a project: how much of the possible clicks does your site get for the observed keywords? 100 would mean first place for every keyword. The tile sits at the top of the “Keywords” tab, with the change versus the previous week and a trend curve.

The calculation uses a published click-through-rate study that shows how often first, second, third place and so on are clicked. If the search volume is known for all keywords, a frequently searched term counts correspondingly more. If it is missing for even one, all keywords are weighted equally, and the tile says so with “weighted without search volume”.

There is one value per calendar week. If you measure twice in one week, the last measurement counts.

The index costs nothing extra, because it calculates with the measurements that exist anyway. It is only comparable within the same project and as long as the same keywords are observed. It cannot be compared with the indices of large search engine tools that calculate over millions of terms.

Usually because the set of keywords has changed. If a keyword has been added, one dropped, or one not measured this week, the new figure can no longer be compared with the old one. The same applies if the weighting changes, for example because a search volume is now available for all keywords.

At this point the curve shows a dashed vertical line; hovering over it with the mouse shows the reason. A change across such a break is deliberately not shown, because a jump that only comes from the new keyword would otherwise look like a success.

A second case: if several weeks lie between two measurements, it says “vs. last measurement” instead of “vs. previous week”.

If you want an unbroken trend, keep a project’s keywords unchanged as far as possible. That is why the weekly measuring run always measures the same ones.

From your own rank measurements, at no extra cost. With every measurement, it is recorded who else is in the top places. For each of these addresses, Aurora Hive Sales calculates the same index over the same keywords.

In the curve, the five strongest businesses appear as lines of their own. Portals and directories are shown separately below, because a large business portal is not a competitor for a business, even if it is at the top for every keyword.

This has a limit: only the first twelve results are recorded. Anyone further down does not appear in the comparison. There, however, the click-through rate hardly counts for anything anyway.

Yes, via the weekly measuring run. It is switched on per project, directly below the “Visibility index” tile, and is off by default.

What applies:

• Only an administration account can switch it on, and only if “Measure ranks for keywords” is on in the settings.
• Every keyword costs one research query per measurement. The preview calculates this before you switch it on: so many research queries per week, roughly four times that per month.
• You decide how many keywords are measured at most. They are always the same ones, so that the trend stays comparable.
• Before every run, it checks whether the monthly budget for research queries is still sufficient for all keywords. If it is not, the measurement waits until there is budget again, usually until the next month.

Without the measuring run, there are only points in the curve when someone measures by hand.

A data query is an order to the data service from which search volume and backlinks come. One query can cover up to 1,000 terms; it is the order that counts, not the number of terms. Rank measurement, map service and market radar, on the other hand, run on research queries, and AI visibility on its own measurements.

How many data queries are possible per month depends on the plan and is stated in your quote or contract. This help deliberately does not give the quantities, for the same reason as with prices: they change, and the help would lag behind.

The current status is shown on the overview in the “Data queries” tile, including the trend over the last few months. In the Visibility settings you can set your own limit under “Data queries per month”. It only takes effect if it is smaller than the plan’s; an empty field means that the plan limit applies.

Once the data queries are used up, the buttons for search volume and backlinks stay greyed out until the end of the month. Everything else keeps running: saved values and snapshots remain visible, and so does the free frequency for the keyword suggestions. With the new month the counter starts again at zero. Anyone who regularly hits the limit should talk to their provider about a larger plan.

The fifth tab of a project. It answers a question that classic search engine evaluations do not ask: are you named when someone asks an answer engine?

More and more people no longer type their question into a search engine, but put it to an assistant and get a finished answer naming a few providers. Anyone who does not appear there does not show up in the whole process — regardless of their place in the classic list of search results.

This is how the tab works:

1. You store questions, the way a customer would ask them — “Who lays parquet flooring in Dresden?”, not “parquet Dresden”.
2. Aurora Hive Sales puts these questions to several answer engines and reads out the answers.
3. It evaluates whether your business is named, in which position, and who is named instead.

The result is a snapshot, not a ranking. Answer engines do not always give the same answer to the same question — that is why the trend over several runs counts, not the result of a single one.

The ones a customer really asks — in full sentences, not as a list of keywords.

What works well:

• The need question: “Who can build me a website in Leipzig?”
• The comparison question: “Which agencies in Saxony work with clubs?”
• The problem question: “My website is slow, who can help?”

What achieves little: your own company name. That an answer engine finds you when asked about you says nothing about whether you reach new customers.

Start with five to ten questions. Each question costs one measurement per answer engine on every run — a long list quickly becomes expensive and is rarely more meaningful.

And keep the questions the same over time. Only then can the trend be compared.

It uses AI measurements — a separate counter, distinct from research queries and data queries.

It is counted by a simple formula: questions × answer engines × repetitions. Ten questions to three answer engines are therefore thirty measurements per run.

How many measurements are included per month depends on the plan and is stated in your quote — this help deliberately gives no quantities. You can see the current status on the overview.

Three things that prevent surprises:

• The area is switched off by default. You switch it on in the settings of the Visibility area.
• Before the run, it says how many measurements it will use.
• Once the quota is used up, the run is rejected instead of being quietly cut short. With the new month the counter starts again at zero; anyone who regularly runs out tops up or changes plan.

Measure monthly, not daily. Measuring more often produces more noise, not more insight.