Using the Instagram API, I collected data via a Python script that crawled and scraped online stores related posts. We learned from my previous article that Indonesia has the most online store posters, so this is a good place to start the investigation. From my collected data, there are more than 6 million online store posts, but do they really coming from online stores? (or user that sells real stuffs?) Or just advertisers that use bot to repost real online store posts? I began my investigation by collecting posts from 4 major cities in Indonesia as samples: Jakarta, Bandung, Jogja and Bali. I chose the cities because of the avid social media users. According to Forbes article, Jakarta is the most active twitter city in the world, and Bandung is in the 6th position.
Instagram online stores, are they real or just bot advertisers?
To get good proportion of the posts, I collected around 30,000 posts from each city. From the posts, I ran another script to collect unique posters from the posts. The result was quite shocking as below. There are only 1,273 unique users in Jakarta, 751 users in Bandung, 1,241 users in Jogja and 574 users in Bali.
Then, I identified which users are advertisers by filtering username that contains “advertise”, “iklan” (means advertise in English) and “promote”. There is only small number of advertiser percentages compare to the real online store. My suspicion on advertisers ruling the online stores world in Instagram seems wrong.
However, I still insist. I ran a script to count number of posts by advertisers on the cities. Now the result shocked me up. Advertisers made more than half of the online store posts! Even more than 75% posts in Jakarta, Jogja and Bali. See below.
Now it is clear that it might be better to collect data from real online stores to perform deeper analyses on Instagram online stores. I was wandering whether the advertisers are bot or not, but I have no idea yet how to identify users as bot. And I’m more interested to answer my next question on Instagram online stores: “Does it really matter for online stores to get more followers?”
Does it really matter for Instagram online stores to get more followers?
Good restaurants usually have more customers to enjoy their good foods. Many visitors coming to Wegmans and Walmart although they would end up buying nothing. In the real world, if we open a store, the more the visitors come, the higher chances customers buying products, the more revenue we can get. Same probably goes with online stores in Instagram. The more online stores being followed, it is likely to get more revenue. However, is this statement really true?
This time I would like to make sure that my data are coming from real online stores. I sorted the collected usernames by number of followers, after removing the advertisers from the list. To get good samples, I ran the script to crawl data from 10 top posters in each city. I removed the usernames that are protected, not open to public. In the end of the crawl, I collected more than 10 thousand posts from each city. It is interesting to see that not all top followed stores made more than a thousand posts. Below visualization depicts number of posts by top followed stores in each city. There seem to be no correlation between number of posts and number of followers.
Then I was wandering about the followers. Are they really looking at the posts made by online stores? Although there is no obvious way to investigate this in Instagram, people will definitely see the posts when they like them. Therefore, like counts can be a good metric. Of course getting likes depends on the content of posts. If the online stores want to improve likes, the first step is to post product pictures that people love. The number of likes received seems to depend on number of followers and how engaged they are. I’m defining the user engagement as number of likes per post. See below for the visualization of the top followers.
In average, the top followers received more than 20 likes per post. Some of them even received more than 70 likes per post. Trendshion, a trendy watch seller in Jakarta even received more than 100 likes per post. It is interesting to see why people love this trendy watch so much. However, we can see that some of the top followers received less than 10 likes per post like Bima Online Shop, a sporty shoes seller in Bandung (when I opened the Instagram of this store, the posted pictures are not interesting). Are their products not good? But why they can have this so many followers? This makes me suspect that they bought followers from somewhere, which are not necessarily real human (can be fake followers).
I also included the number of comments per post, but it is found that the numbers are very less. It seems when people see products sold in Instagram, they contacted the sellers directly from the phone number and Whatsapp chat that are usually listed under their Instagram bio.
In conclusion, number of followers does not guarantee people from liking the photos. While it might be true that great number of followers increase the chances of products being seen, but good products matter a lot.
See the scripts I used for this project here: https://github.com/girikuncoro/shopinstagram2
Note: I made the visualizations using D3.js. Look at under “viz” folder for my visualization script.