Reading ranking metrics: best position, average and page-1 share

Best position: optimistic and narrow

The best position is the highest place a product reaches for any of its search terms. It is easy to understand and works as a headline: "position 2 on otto.de".

It hides breadth, though. A product at position 2 for its brand name and invisible for every generic term has the same best position as a product that sits near the top everywhere. Brand terms in particular pull this metric up; see brand keywords vs generic search terms.

Average position: depends on "not found"

The average across all search terms looks fair but leaves one question open: what happens to terms for which the product does not appear in the pages checked? There are three common approaches:

• Leave them out: the average only covers terms that were found. It looks good even when half the terms are missing.
• Fixed substitute value: not found counts as the first place after the checked range, for example 361 when three pages of 120 products each are checked. More honest, but a single missing term shifts the value a lot.
• Report separately: average of the found terms plus the number of missing ones. Usually the most readable option.

Pick one approach and stick to it. An average whose calculation changes between two reports is not comparable.

Median, page-1 share and count with a position

• Median: the middle value when you sort all positions. A single outlier, say position 300, barely distorts it. It describes the "typical" keyword better than the average.
• Page-1 share: how many of the tracked terms are on page 1, which on otto.de means within the first 120 products (as of October 2026). It shows breadth and is easy to explain. Why page 1 is that large is covered in page 1 on otto.de.
• Count with a position: for how many terms the product is found at all. When this number rises, the product is becoming relevant for more searches, even if the positions are still far back.

Also decide whether you calculate with the overall position or the organic position. Ads shift the overall position; more in overall, organic and ad position.

Worked example: one product, five search terms

Example with made-up numbers, organic position, three pages checked (positions 1 to 360):

• Brand term: position 2
• Product type plus size: position 9
• Product type plus material: position 47
• Product type alone: position 160
• Long-tail term: not found

This gives:

• Best position: 2
• Average of found terms: (2 + 9 + 47 + 160) / 4 = 54.5, plus 1 term not found
• Average with substitute value 361: (2 + 9 + 47 + 160 + 361) / 5 = 115.8
• Median (not found as the last value): 47
• Page-1 share: 3 of 5 = 60%
• Count with a position: 4 of 5

So the same product is "position 2", "average 54.5" or "average 115.8" depending on the metric. None of them is wrong. For optimisation, the interesting term here is the one at position 160: it is the head term, it sits on page 2, and a move to page 1 would lift the page-1 share to 80%.

Which metric for which decision

• Optimising a product: look at the individual terms, not a total. The metrics only show you where to look.
• Reporting to management or clients: page-1 share and count with a position, plus the median. They are stable and easy to explain. What such a report looks like is described in the monthly ranking report.
• Alerts: a falling page-1 share, or a term that switches from found to not found. Both are clearer than the average moving by a few places.
• Promoting your own product: the best position, but only together with the term.

How to keep the numbers comparable

1. The same keyword set over time. Adding or removing terms changes every metric without any change in the ranking.
2. The same treatment of "not found" in every report.
3. The same type of position: organic or overall, not mixed.
4. The same check frequency, daily or weekly. MarktRank keeps the full history, and the CSV export contains the positions per search term, so you can recalculate any of these metrics yourself.

Frequently asked questions