There is an additional type of traffic from image search in the report, should this type of traffic be handled separately for content?
There is a type of search by image data appearing in the search console report. Should the site create separate content for it? No need to open a separate content line for it, but we should first do a consistency check: whether the description of the image itself, the main topic of the page where the image is located, and the questions to be answered on the landing page all point to the same thing. Most of this traffic is not a new issue, but an old issue with a new trigger
What kind of queries is this type of statistics
A new filtering dimension groups search results that use images as part of the query separately, that is, users search with images or initiate queries in the mixed image and text entry. It counts the display and click sources of search results, not a new page type on the site side.
This means that the data of the original web report has not disappeared, but has been split into two finer groups: conventional text queries and multimodal queries. The total amount remains the same, and when looking at trends, the new dimension needs to be cut back.
What is the relationship with the original web report?
| Dimension | Conventional web queries | Queries initiated with images | Key points on the site side |
|---|---|---|---|
| Query expression | Keyword text | Image or image-text mix | The landing page theme should be inferred from the image |
| User intent | Clear information requirements | Find similar, find source, find similar | Description text is more important than rhetoric |
| Optimization handle | Title matches the main text | Image description and structured text | Both use the same batch of keyword libraries |
| Data visibility | Long-term comparable | Incomplete gray period | Record the baseline first and then discuss the rise and fall |
From late September, this type of multimodal began global gray, and the time when this data appeared at different sites was not consistent. During the gray period, do not judge the traffic decline because some data is not visible, first record the baseline of the current window, and then compare when the口径 is stable.
Should we look at the picture description first or the landing page first
First, look at the image description, then view the landing page. The image description and alternative text are the most direct matching signals in this type of query. If the image file name and description are something like 'IMG_001', it's as if you've abandoned this step. The main theme of the landing page determines whether it can accommodate after being matched.
Check the order and suggest three steps: First, align the description text of the site images with the theme of the column; second, remove decorative images unrelated to the text; third, let the same theme image album be concentrated on a page that can be described, rather than scattered in multiple articles.
How to supplement this type of query with a keyword library
The value of keyword library management is directly reflected here. The query words for image search are often generated by users after the event, such as style names, model numbers, and scene words, which are usually not in the original word list sorted by text queries. The method is to supplement these query words back into the keyword library from the reports, and then check whether there are corresponding landing pages; add pages for missing pages, and modify the description for pages with descriptions but not adequate.
After completing it, aggregate it by column to avoid the situation where the same word corresponds to dozens of unrelated articles.
The position of automatic image matching in this type of traffic.
The automatic title image allocation capability can automatically allocate images based on the title content, saving the time for image allocation during mass publishing. However, it solves the problem of 'having images', but not 'whether the images are correct'. The matching quality of image search depends on whether the information carried by the image is consistent with the query, and the images selected by the machine based on the title are easy to be out of place with the actual objects discussed in the text.
Therefore, the division of labor is recommended as follows: automatic image allocation is responsible for the allocation of images for the first draft and low-priority content, and manual verification is only required for the core landing pages and product category pages; after the description is changed, the exposure changes of such pages should be recorded separately, as the input for the next round of keyword library supplementation.
Common questions
Should we create a new page specifically for image search?First, make sure the image descriptions on the existing pages are matched with the keyword library. Most sites can achieve this kind of traffic growth at this step.
Can this data see the specific query words?The report is grouped by search type, specific words are still subject to aggregation and privacy restrictions, and trend judgment is given priority over word-by-word analysis.
When does the gray period end?The global gray period started in late September, and the launch time of different sites is inconsistent. It is recommended to draw conclusions after two consecutive comparable periods.