What can you do for my company?
This was the initial client’s question, they also gave us a series of concepts that they considered essential for their business:
Search for incluencers through both opinion and engagement leadership detection in order to address them to:
The following is the first phase of the work developed by Network Outsight, consisting in a first assessment and provisional conslusions. The data shown correspond then to a preliminary analysis that served as the beginning og the moniroting strategy with a stable nature.
First step: 9 days of conversations on networks were initially monitored.
|Search term||Scrapbooking, pintura decorativa, silueta dm, decoupage. Language: Spanish|
|Monitoring period||05/06/2015 – 14/06/2015|
|Captured tweets||511 out of 300 accounts|
|Captured mentions||321 out of 130 sender and 124 receiver accounts|
Second step: A recount for the search for the most commonly used and associated words by the users on Twitter was carried out. This revealed the importance of some associated concepts that were not initially taken into consideration. In the wordcloud bellow the 100 most common concepts of the analysed database can be observed, the size of each word is determined by the users’ use, being the most used words bigger than the less used:
The client provided us with the most relevant words of the sector according to his opinion. But he did not give us two essential words for the networks users that make crafts such as ‘sellos’ (= seals), ‘troqueles’ (= dies), or the anglicisms ‘craft’, ‘handmade’ or ‘DIY’ (Do It Yourself). They are concepts that involve very important communities in this sector. The client was then underestimating the opportunity to address a very large audience of potential customers that were easy to approach.
Third step: We decided to also analyse the messages flow in order to make decisions on when to publish. There are lots of automatic tools to detect the publication hours of the users of an account; we were more interesed in the craft fans’ timetables though, even if they did not follow our client.
The company’s owner, who personally conducted its social networks, thought that his activity was primarily a hobby taking into account the type of customers he had. That is why he had been tweeting from the very beginning only on weekends. ‘If it was a hobby, customers will make crafts on the days they do not work.’ And if this was like this, they would make crafts on weekends…, but they did not tell what they had made and the problems they had had during the week.
On weekends they made crafts, but they talked about it during the week. In fact, the days on which they spoke less were Satudays and Sundays (5, 6, and 13 June). Since Twitter is a very talkative social network, messages are mostly instantly read, and the owner of this company tweeted and nobody paid attention to him. A tweet made on Saturday and Sunday does not survive on Monday, and less the rest of days.
Fourth step: We started to look for the influencers of the conversation using Social Network Analysis. There is bellow the monitored conversation in a network graph; it has been anonymised substituting the Twitter accounts by numbers. Each node represents a Twitter account and each edge (line) between nodes represents a mention relation (@user). The thickest edge in the network (from node 9 to node 74) represents a total number of 12 mentions.
The nodes have been anonymised and presented as numbers. When clicking on each node, some of its structural properties can be learnt. In this case the Weighted Incoming Grade (the number of received mentions), the Weighted Outgoing Grade (the number of mentions made), and the Engagement Centrality (the number of edges between the nodes in the graph among which a specific node is found) allow us to find opinion and sharing leaders. For instance, the node 74, related to a blog on craft present on Twitter, is obviously an opinion leader, because he receives lots of attention. However, the most powerful node in engagement teems is the node 9, someone passionate about Scrapbooking who is a key sharer of 74’s messages and some others. Without the node 9, the maroon subcommunity in which nodes 45 and 131 are located would not have access to the blue central community.
In the graph there are a total number of 222 nodes and 321 edges, 56 of which represent more than one mention. There are 44 different components (groups of nodes isolated from the rest), and 49 communities (groups of nodes densely connected to one another) that have been highlighted with different colours. All this shows that the captured conversations have been carried out separately, without any significative interconnector.
Among the influencers of the conversation, a bloggers community constituted in the network stands out, as well as a radio speaker with a very big influence on his more than 50k followers. These influencers can be used to share our message in a very large network constituted around these located accounts, and make real commercial and marketing campaigns, such as, send them a craft, invite them to the factory, interact with them, etc.
Perfoming commecial campaigns for customer acquisition
With all this information and knowledge of the sector and the market, of the network users that talk about crafts, of the influencers and theis importance on the network, and so on we can arrange strategic actions like Twitter campaigns. One of the biggest advantages of making campaigns on Twitter, targeting from the data obtained, is that addressing 20-30k accounts previously selected and with an apropriate message to their hobby can have a total cost of only 20€.
If you wish more information on the implementation of all this work into your company, brand or product, contact us: