PDF Publication Title:
Text from PDF Page: 005
(iv) Follow/Unfollow and (v) Comment activities. Being more precise, for what concerns the unfollow operations, the Social Bot will unfollow only who does not follows the given user. Instead, for what concerns the likes and the comments, it works randomly on the basis of the retrieved posts. The Bot operations can be steered through a list of param- eters. The most relevant instabot-py parameters for us are the instagram account (i) username and (ii) password, the number of (iii) likes and (iv) comments that the Social Bot must leave every day, the number of (v) following/unfollowing operations to do every day and the (vi) hashtag and localization list. Hashtags and locations are treated in a similar way, with the only difference that the locations have the prefix ”l:”. In particular, our Social Bot ended its activities doing 1000 (1000) like, 24 (400) comments, 300 (500) following and 500 (500) unfollowing operations. Instead for what concerns the hashtag we used the most relevant photography hashtags and we used ad locations the main Emilia-Romagna cities. Between parenthesis are listed the respective limits to do not pass to decrease the possibility to obtain an activity suspension by Instagram [17]. However, we do not started directly from the reported parameter values, but we slowly increased those value until we reach the indicated values. Moreover, many times the configuration parameters were changed, to increase or decrease the Social Bot activities, to fit better to certain post themes, to improve the hashtag list quality or the Bot targets. In this way we were able to manage better the current needs (i.e. a number of following too high or too low, avoid the risk to be suspended by Instagram). With this approach, and as our experimental result will demonstrate, it is possible to overcome the Instagram algo- rithm limitations. This because the Bot will interact with a lot of users, that with a good probability could visit the user profile, leave some feedbacks (i.e. likes or comments) or follow the profile. However, the instabot-py project [14] were not satisfying all our needs. Thus, to better fit the introduced strategy, we decided to extend the project adding: IV. EXPERIMENTAL RESULTS To have a good evaluation of the Social Bot activities, we started taking a snapshot of the most relevant profile statistics achieved before the use of the Social Bot deployment and then we repeated the same process after the Social Bot utilization. In particular, during our analysis, we first focused our attention in the following profile information before starting the Social Bot (Figure 1): Figure 1. Growth Results with normal activity From Figure 1, we can retrieve that, even after 22 months of activity, the profile did not gain a good success. Indeed, the follower metric was not satisfied. Coherently with what rated in [13], we think that these bad results are driven by the fact that the posts were not spread correctly to new users by the Instagram algorithm. Indeed, the were mimiking an user focused on providing a good post content without being concerned on how the Instagram algorithm is working or on how to exploit it in the best way, as many of the Instagram users are doing. The fact that the @PROFILEA posts were not well spread over Instagram is proved by the maximum post reachability value, that is comparable with the number of followers (550 vs 541). So, considering that, since most of the user followers will see all the followed profile posts, the probability that the post reaches new users is very low (<20%, 15% in our case), and this will affect all the other statistics. Being the reachability an indicator for how many account the user posts can reach, having a low reachability has as direct consequence, even if the user posts good contents, to have a low probability to receive likes, comments or new followers. Those considerations are proved by the weekly profile visitors statistics. Indeed, only 100 users by week, both new users or followers, were visiting the profile, certifying that, even if the user intent was to reach more new users as possible, something was wrong in his strategy. Another aspect proving that the user strategy was not working, is the ratio followers/posts is equal to 4.15, really low with respect with the Instagram possibilities and expectations. In cases of good growth, depending on the posts frequency, the followers/posts ratio can be greater than 100, a big difference from the actual situation. This situation also affects the likes result that, in average and after almost 2 years of activity, • • • Separation between the Bot and User Space: in this way the Social Bots activities do not affect the user activities. This were implemented creating a new DB table in which record if a certain user were added from the bot of from the user; Targeting on influencer followers (new parameter: profiles target list parameter): this extension allows to apply the follow/unfollow strategy on profiles similar to the own profile, increasing the probability to obtain likes, comments or new followers, since we already know that the followed user is interested on similar posts (4th point of the Follow/Unfollow classification); Follow/Unfollow on a give influencer list: do a fol- low/unfollow activity on an influencer helps to exploit the same principle of the previous point, but being followed from other users that are applying the follow/unfollow strategy (5th point of the Follow/Unfollow classification).PDF Image | Analysis of Growth Strategies in Social Media: The Instagram Use Case
PDF Search Title:
Analysis of Growth Strategies in Social Media: The Instagram Use CaseOriginal File Name Searched:
Analysis_of_Growth_Strategies_in_Social_Media__the_Instagram_Use_Case_FINAL.pdfDIY PDF Search: Google It | Yahoo | Bing
Cruise Ship Reviews | Luxury Resort | Jet | Yacht | and Travel Tech More Info
Cruising Review Topics and Articles More Info
Software based on Filemaker for the travel industry More Info
The Burgenstock Resort: Reviews on CruisingReview website... More Info
Resort Reviews: World Class resorts... More Info
The Riffelalp Resort: Reviews on CruisingReview website... More Info
| CONTACT TEL: 608-238-6001 Email: greg@cruisingreview.com | RSS | AMP |