First, I'd try to narrow the problem. With Google Images I assume we are talking about the images tab within Google Search. Therefore, the answer's scope is limited to design the experience of typing a query and receiving images related to the query as a result.
Secondly, I'd think about potential groups of users and motivators/pain points. This is a challenging phase as searching for images is a very broad task. Instead of segmenting users by "meta personas" I'd segment them in 2 broad groups differentiated by the expectation of the result.
- Users who already have a clear idea of what they are expecting: This could be academic people or professionals. People who look for images in order to use them in a professional or illustrative context a companion resource for a complex idea. Motivators could be including the images in a pitch deck, article, publication, etc. It is somebody who is already close to the concept queried and is looking for a graphical representation of this known concept.
- Explorers, those who are looking for the visual represnetation of a concept for the first time: This could be students, children, or people who were reccently introduced to a new concept.Users who look for a better sense of how does the world looks like. Here the motivation is more research related, perhaps the user does not have a very clear idea of what to expect prior to the query.
We could also consider as a potential third group, people who are looking for information or knowledge within an image. For example, mind maps or visual representation of flux diagrams (even memes). Here it's not only relevant the visual aspect of image but also the embedded information within the image.
We could now think of use cases and pain points for each of the 3 main user groups previously mentioned:
- Group 1 is looking to find an image they already have in their minds. High quality and visual impact is valuable for them.
- Group 2 is looking to be surprised, to visualize an unfamiliar concept. High variance of results could be more helpful for them in order to develop a complex, un biased, idea of how does something looks like
- Group 3 is looking for high quality content within the image. visual quality is not as relevant as content accuracy and relevancy related to the typed query.
Now, taking into consideration that these 3 groups are currently aiming to solve their very different goals with the same product by performing the same action: typing words and hoping to get the expected result, we can derife a couple of main features:
- "academic or professional search mode" : in which imaged displayed are highly illustrative and copyright free, ready to be used in documents
- "explorer mode": in which Google could even tell small snippets of stories by showing illustrative, variable images of complex concepts
- "Search for concepts": in which you could tell google if your query refers to an object within the image or information/text within the image
All 3 mentioned features to design refer to the whay in which users could use the actual product. However, sometimes users whant to use the same search power for their own ontent or for a specific subset of the website. Having the chance of choosing the source in which the search qill be performed "EG: the whole web, owned pictures, professional galleries" could help solve the problem os niche-specific users.
Wrapping it up, "Google images" or images search is a very broad product with the flexibility to solve for many. However, by adding customizable features such as specific search modes or specific data sets to search into, the exprience could be enhanced in order to do a more granular image search, if the user needs require it.
