Research on Google Lens rankings has examined page titles, URLs, image filenames, surrounding text, and image placement. The links between rankings and alt text or responsive images are less clear. However, the main data comes from a study published by Backlinko in 2020 and describes search results at that time. It does not prove that these characteristics affect rankings and should not be treated as a set of current algorithm rules.
Titles, URLs, and filenames: matching keywords offer clues
The study analyzed 65,388 Google Lens results. It first used the Google Vision API to generate text labels for the images used to search, then checked whether the text on result pages matched those labels. For example, an image of a cat might be labeled “cat”; this does not mean a user actually entered that search term.
- 32.5% of result pages had a title containing a keyword that matched the search image’s label.
- 29.9% of result pages had a URL containing a related keyword.
- 22.6% of result images had a filename that matched the visual search.
Based on these findings, the authors speculated that matching text might offer a ranking advantage. However, these percentages do not indicate how much rankings might improve after changing titles, URLs, or filenames. They are also insufficient to establish each attribute’s ranking weight.
Surrounding text provides context, not a word-count target
The result pages in the study contained an average of 1,631 words—not 1,631 Traditional Chinese characters. Citing Google’s advice at the time to place images near relevant text, the authors suggested that surrounding text might help explain an image’s content and context.
This supports making text and images work together, not requiring every page to reach a fixed length. The study does not establish that increasing the word count itself improves Lens rankings, nor does it identify an optimal length.
Image placement offers clues; responsive design showed no strong link to rankings
33.1% of result images appeared in the top quarter of a web page. The authors considered this consistent with Google’s statements at the time about prioritizing images higher up on a page. However, this does not guarantee that moving an image to the top will improve its current Lens ranking.
Meanwhile, approximately 13.1% of result images used responsive design, and the study found no strong relationship between responsive design and visual search rankings. This does not mean responsive images have no value: the study still notes that they may improve the user experience, but that benefit does not imply a direct boost to Lens rankings.
Alt text: limited ranking evidence does not mean it should be omitted
Only 11.4% of result images had alt text matching the best-guess label for the search image. The authors of the 2020 study therefore speculated that alt text might play a smaller role in Lens optimization, while explicitly stating that they did not know whether Lens used it.
Expert advice published on Neil Patel’s website in 2024 still emphasizes descriptive alt text, file compression, and using images alongside content to help readers understand it. These are broader image SEO recommendations. They do not disprove the study above or establish how much each practice might improve Lens rankings.
Google Lens ranking factors: perspectives from across the web
Are the “keywords” in visual search research text entered by users?
Not necessarily. Backlinko’s study used the Google Vision API to generate text labels for search images, then checked those labels against titles and other content on result pages. Thus, a text match refers to a match with a tool-generated image label; it cannot be taken to mean that a user entered that term. This also limits direct comparisons between the study’s findings and conventional text search data.
Does alt text that does not contain the image label count as incorrect?
Not necessarily. A label mismatch does not mean the description is wrong. The 2020 study only checked whether alt text matched the best-guess label for the search image; it did not check the accuracy of each description that failed to match. The study’s authors were cautious about alt text’s role in Lens, while a separate article offering expert advice in 2024 still emphasized descriptive alt text. The two sources cover different ground, so a label match alone should not determine whether to keep alt text.
Should the priority be making pages longer or making images and text work together?
The research and expert advice provide more support for pairing images with relevant text than for pursuing a fixed length. The 2020 study observed that result pages contained substantial amounts of text, and the authors suggested that text surrounding an image might help provide context. The 2024 expert advice likewise emphasizes using images alongside content to help readers understand it. These are observations and recommendations, not proof that lengthening an article or integrating images into the content will necessarily improve Lens rankings.
References
- Google Lens Study: Results From 65,388 Visual Searchesbacklinko.com · 2020-06-09T09:03:49-05:00 — Discusses: Are the “keywords” in visual search research text entered by users?; Does alt text that does not contain the image label count as incorrect?
- Image SEO: Tips for Image Optimizationneilpatel.com · 2024-11-26T20:00:00+00:00 — Discusses: Does alt text that does not contain the image label count as incorrect?; Should the priority be making pages longer or making images and text work together?
