Haha that went right over my head @StevenBerlin. If i were an algorithm I wouldn’t miss things like that 
Yes apparently that’s true @WilfriedB and very astute of you to think out of the box on that one!
1 Like
Okay. I’ve come back and voted B. The details in Photo A are indistinct due to the distance away.
1 Like
Good to see that you have voted @tony_b because @Trail_blazer and I’m sure others, are keen to wrap this up.
2 Likes
Dear Adam,
Similarly to my post about choosing between the three supermarket images (to which I replied earlier), analyzing the two photographs above, the one that has the highest probability of receiving more views on Google Maps is photo B (daytime photo).
The reasons are:
-
Clarity and Detail:
Daylight allows for clearer and more detailed images. Users can clearly see the entrance, windows, signs and other important elements of the supermarket.
This level of detail is essential for users who are looking for practical information, such as the appearance of the store, opening hours and accessibility. -
Realistic Representation:
Daytime photos tend to represent the appearance of the store during opening hours, which are the hours of greatest interest to most users.
This helps users get a precise idea of the place and orient themselves more easily. -
Visibility and Usability:
Daytime photos are generally brighter and more visible, which makes them more attractive to users scrolling through images on Google Maps.
Users are able to see details better. -
Practical information:
People who search for a supermarket on Google maps do so mainly during the day. Having a photo taken during that time allows customers to have more useful information.
While the night photo can create a suggestive atmosphere, and aesthetically it could be more pleasing, the daytime photo offers greater utility and visibility for Google Maps users.
Best regards from Rome, Italy
2 Likes
Thanks Alex for your detailed reasoning. Well we will all know very soon how the votes ended up in each of the 3 polls in the trilogy and the finishing order in terms of most views of the 3 storefront photos.
1 Like
If anyone read my previous reply, regarding the pair of night shots, one from an angle and the other straight on, this answer is also along the same lines of thought. Below has a larger amount of pixels representing the business sign, there’s a nice sky with clouds. Both photos in this case do NOT show the wording on the side of the building, and there are no license plates that the algo would throw out. - Steve
I like Photo-B. It was taken in daylight. The sky above the store looks silver. You can see the ALDI sign clearly. The design on the wall behind the sign is bright. There are no car shadows in front. It’s a close-up picture. The daylight and lights inside the store are nice. However, the views might depend on where it shows up on Google Maps.
1 Like
Hello everyone
Here is my attempt at collating and summarising the discussion on this second part of the trilogy which like the first part, has been incredibly insightful, revealing a diverse range of perspectives on which of the two straight on storefront photos, one taken during the day and the other at night, has the most views.
Overall Sentiment & Initial Thoughts:
-
Photo A (Nighttime) Supporters:
- @StevenBerlin initially strongly favored Photo A due to perceived sharpness, better alignment with the rule of thirds for the main ALDI sign, and the visibility of the smaller signs on the door (which he believed were obscured in Photo B). He doesn’t believe day or night influences the algorithm.
- @HasanBTB simply voted for Photo A without detailed reasoning initially. Later mentioned it looked “epic & different.”
- @Peadar found the “ghost from the past” (referring to the previous night photo) better, noting Photo A shows the whole storefront, liked the roof lighting, and felt it looked more appealing if hungry. Ultimately predicted Photo A would have more views across both rounds.
-
Photo B (Daytime) Supporters:
- @JanVanHaver voted for Photo B based on a “gut feeling” that the algorithm prefers daytime photos.
- @Annaelisa favored Photo B because it appeared clearer and closer to the business, directly contradicting @StevenBerlin’s sharpness observation.
- @TerryPG generally prefers daytime photos and initially leaned towards Photo B but predicted the algorithm would favor the nighttime Photo A.
- @tony_b ultimately voted for Photo B, citing the indistinct details in Photo A due to distance.
- @shunsader leaned towards Photo B due to a larger pixel representation of the business sign and the presence of a “nice sky with clouds.” Noted the absence of side wording and license plates in both.
- @PrasadVR explicitly stated liking Photo B, highlighting the daylight, “silver” sky, clear ALDI sign, bright wall design, absence of car shadows, close-up view, and the nice balance of daylight and interior lights. Acknowledged view count might depend on placement on Google Maps.
- @AT_Rome provided a detailed rationale for choosing Photo B, emphasizing
(1) Clarity and Detail: Daylight provides more visible essential elements,
(2) Realistic Representation: Daytime reflects opening hours, the time of most interest,
(3) Visibility and Usability: Brighter and more attractive for scrolling users and
(4) Practical Information: More useful for users searching during the day.
-
Undecided/Cautious:
- @Trail_blazer initially focused on the importance of clear, sharp, and well-lit logos in night photos. Didn’t make a definitive vote but emphasized clarity and details. Later acknowledged seeing preferred night shots but questioned if it applies to supermarkets.
Key Reasons and Arguments Presented:
-
Signage Visibility & Size:
- @StevenBerlin initially focused on the visibility of smaller door signs in Photo A and the rule of thirds alignment of the main sign. Later emphasized the importance of visible and in-focus store signs for the algorithm, even suggesting it looks for multiple signs.
- I countered @StevenBerlin by pointing out the larger size of the ALDI sign in Photo B and questioning if users would focus on minor signage details.
- @Rednewt74 measured the pixel size of the ALDI signs, noting the significantly larger sign in Photo B, suggesting this would be an advantage on phone screens.
-
Image Quality (Sharpness & Clarity):
- @StevenBerlin perceived Photo A as sharper and more in focus.
- @Annaelisa found Photo B to be clearer.
- @Rednewt74 agreed with @StevenBerlin that Photo A appeared slightly clearer.
-
Daytime vs. Nighttime Preference:
- @Trail_blazer suggested night shots can be preferred for certain businesses (restaurants, bars) due to the “vibe” but questioned its applicability to supermarkets.
- @JanVanHaver believed the algorithm generally favors daytime photos based on intuition.
- @TerryPG stated a general preference for daytime photos.
- @StevenBerlin explicitly stated he does not believe day or night influences photo featuring.
- @Rednewt74 thought the algorithm generally prefers daytime photos but acknowledged night shots can have better exposure balance in some cases, though not for this supermarket.
-
Rule of Thirds:
- @StevenBerlin felt the ALDI sign in Photo A was better aligned with the rule of thirds.
- I questioned if the signs in both photos weren’t similarly aligned.
- @Rednewt74 observed that neither sign fell directly on the rule of thirds points but were similarly positioned on the top third line.
-
Composition & Framing:
- @Peadar noted Photo A shows the whole storefront, while Photo B doesn’t, making him wonder what’s on either side. He did like the leading lines of the cars in Photo B.
- @Rednewt74 pointed out that Photo A has a larger area of the sky and more of the parking area in the lower third, while Photo B focuses more on the building in the upper third.
-
Algorithm Behavior & Learning:
- @tony_b suspected that this experiment was designed to reveal unexpected algorithm behavior.
- I emphasized the goal of understanding the AI’s preferences to improve the quality of our photo contributions.
- @StevenBerlin agreed with the aim of understanding the algorithm to improve future photography.
- @Rednewt74 shared insights he gained from Gemini about how AI analyzes images by recognizing patterns of shapes, edges, colors, and light, aiming to identify helpful and clickable images. He also expressed skepticism about the algorithm’s current effectiveness.
-
Impact of Multiple Photos & Ranking:
- @Trail_blazer shared experiences of the initially top photo slowing down and being overtaken by others uploaded later.
- @tony_b corroborated @Trail_blazer’s observation with a personal experience of a secondary photo eventually becoming the top photo.
- I myself confirmed that I too had observed this phenomenon and also mentioned that the photos in this poll were deliberately taken and monitored over a year.
- @Peadar wondered about the importance of the first photo uploaded to a POI and the contributor’s overall “value” to the AI.
-
Irrelevant Content & Algorithm Flaws:
- @Rednewt74 highlighted the algorithm’s apparent inability to filter out irrelevant content like personal photos or TikTok videos.
- @Trail_blazer echoed this concern, mentioning irrelevant videos getting good placement and questioning if Google addresses copyright issues with such content.
-
Local Guide Experience & Trustworthiness:
- @Peadar questioned if the length of time as a Local Guide and their trustworthiness influence the AI’s preference.
Key Themes and Points Raised:
- Reinforcement of Daytime Preference: Several Local Guides explicitly favored the daytime photo (B) due to clarity, detail, and its representation of the store during typical operating hours.
- Image Detail and Visibility: The lack of clear details in the nighttime photo (A) due to distance was a deciding factor for @tony_b
- Signage Size (Pixel Count): @shunsader quantified a reason for preferring B based on the larger pixel representation of the ALDI sign.
- Compositional Elements: @shunsader also noted the “nice sky with clouds” in Photo B as a positive attribute.
- Absence of Negatives: The lack of license plates (mentioned by @shunsader) continued to be seen as a potential positive for both photos.
- Algorithm Focus on Utility: I myself and @AT_Rome echoed the idea that the algorithm prioritizes helpful and informative photos that users are likely to click on, rather than purely aesthetic appeal.
- POI Influence on Views: @PrasadVR astutely pointed out that the photo’s placement within the Google Maps listing could significantly impact its view count, regardless of its inherent qualities.
- Discussion on AI’s “Vision”: I highlighted @Rednewt74’s shared Gemini chat, emphasizing that AI analyzes numerical data representing the image rather than “seeing” it like a human.
- Skepticism Regarding Business-Uploaded Photos: I expressed reservations about the consistent prominence of potentially lower-quality business-uploaded photos.
In Summary:
The discussion in Part #2 of this trilogy reveals a variety of opinions and reasoning regarding which photo (daytime straight-on vs. nighttime straight-on) might have more views. Factors considered include signage visibility and size, image sharpness, the perceived preference of the algorithm for day or night, rule of thirds, composition, and the overall goal of creating helpful and clickable images for Maps users. For some there was a clear leaning towards the daytime photo (B) based on practical factors like clarity, detail, and its relevance to users searching during the day.
The discussion continued to explore the potential criteria the Google Maps algorithm might use, balancing aspects like signage, composition, and overall utility. The limitations and sometimes seemingly illogical outcomes of the algorithm’s choices remained a point of discussion. Several Local Guides also shared their experiences and theories about the algorithm’s behavior regarding multiple photos, ranking, and its potential flaws in identifying irrelevant content. The conversation highlights the ongoing efforts of Local Guides to understand the often-opaque logic behind Google Maps’ photo selection and ranking processes.
You will appreciate that in a detailed discussion like this, it isn’t easy to collate and summarize everything 100% correctly so if you find that I’ve missed something important or misinterpreted anything you have said, do let me know. In any event, we’re now clearly moving towards the conclusion, with anticipation for the final results of all three polls.
8 Likes