Prove/disprove of microphone used for targeting Ads
Download as PPTX, PDF1 like1,642 views
The project aimed to determine if ads in apps like Facebook, Instagram, and Snapchat are influenced by microphone usage. Initial findings suggest Snapchat explicitly states they use the microphone for ads, while Facebook and Instagram do not clarify this. However, various marketing strategies and user behaviors also likely contribute to the targeted ads observed, indicating other factors beyond microphone listening may be at play.
Prove/disprove of microphone used for targeting Ads
- 1. Prove/disprove of Microphone used for targeting ads Made by Daksh Raj Chopra (40054446) Manav Chawla(40058312) Shabadpreet Singh(40076462) Atul Sharma(40070401) Olufunsho Laniyan (400568986)
- 2. Explanation of the project This project was based on the observations from our ads coming in the app. We were supposed to find if these come from other source, microphone too. We had to create an experiment which through we can completely conclude if our device microphone listens to our chat. We started with reading the Privacy Policies for all three apps i.e. Facebook, Instagram and Snapchat. Then choosing our keywords to see if that category/product shows up in the chat. Conclusion: Lets wait for the results.
- 3. Privacy policy Starting with this we read all the term and conditions for the three app our experiment was based on. 1. On the right, we have the screenshot for the Privacy Policy for Snapchat, which clearly states that they use our microphone for our ads.
- 4. Privacy policy(Cont.) 2. But in case of Instagram and Facebook(both owned by same company), there was no mention if they are using our Microphone.
- 5. Ideal world • Best approach: To choose New hardware, create new accounts and start test. • But, creating new account is another problem, since psychologically we will end searching the same categories you like and creating similar ids as before(same version of our self). • Ads are not just linked to app. They are also linked to device(Mobile Advertiser IDs).
- 6. Framework 1 for 30 days(Pseudo Algorithm) 1. Collect ads for 10 days. 2. Combine the list of all the ads (L). 3. Select 2x,2y s.t 2x,2y ∉ L (where x is a category and y is a product). 4. For (x1,y1) open Facebook, Instagram and snapchat and start talking. 5. For (x2,y2) unlock phone and start talking.(background). 6. Repeat Step 4 and 5 till 15 days.
- 7. Framework 1(additional) • We occasionally met to share ads(the old fashion way) to prevent advertisements(data leak through keyboard or search). • Choosing ads that does not belong to common list, makes our device act like new. • ads are choosen not just on the basis of list also about things that we lack interest in.
- 8. Support framework for next 30 days • Collect ads for 10 days. • Combine the list of all the ads (L). • Select 3x from L such that • It is not one of your collected ad • It is something that we have never “intuitively” searched or seen before For (x1,x2,x3) open Facebook, Instagram and snapchat and start talking. Repeat till 15 days
- 9. Why support framework • It considers into fact the pay per click rate.(The ads that we are currently seeing are highly paid) • For the case of Framework 1, the assumption is that all ads are equally likely to appear(say If I choose cheese and you choose car, chances are same). • We cannot deduce just on the basis of support framework because On Facebook, the targeted ads change based on what your friends are doing. Also, snapchat requests permission to contact list. Ads based on friends interest is also very likely.
- 10. Screenshots and result • Mattress (background) • Bag(background) • Electric vehicle(foreground) • Oneplus(foreground) • Goat(foreground) • Second cup(got Tim Hortons) (foreground) • Games(sentences with games)(foreground) • Canada goose(foreground)
- 12. Bag(background)
- 15. Goat(foreground)
- 16. Second Cup(got Tim Hortons) (foreground)
- 18. Canada goose
- 19. Other result There were some ads which were not based on experiment but came since we talking about them a lot. • Jack and Jones • Canada goose (got some other jacket) • Subaru • Fizz
- 20. Jack and Jones
- 21. Canada goose
- 22. Subaru
- 23. Fizz
- 24. Conclusion Well from the ads given above it seems somehow that yes these Apps are listening to our chats all day and giving the ads accordingly. But there are other several reasons are for the appearance of these ads like Geographic Target Marketing, Gender-Specific Marketing, Income- Sensitive Marketing, Age Target Marketing, based on purchase history and many more. The reason of our result(ads) could be because of one of the above. Hence saying these ads came only because we were talking about it could be inconsistent.