Social Media Analysis Software: Proposal & Business case.

A software application / suite that lets you keep track of and analyse a subset of all social media posts, and analyze correlations / relationships between them.

Has a rather different focus from existing web hosted applications (more details below).

Identification of End Users (types / groups), and their requirements.

Considerations:

– The average human being cannot program.

– The average human being cannot maintain a database (well).

– The average software developer cannot be bothered to go through the various different (and unique) API’s for each social network.

– Aforementioned API’s are mostly for web apps, not local apps.

– Which require logged in access.

– And rate limit.

– And require subscription.

– And are subject to change.

– And you still don’t necessarily possess the data.

Casual social network users:

– I would like to keep track of me and my mates.

– I would like to keep a copy of what they post.

– I would like to know who comments on whose posts, and how often.

– I would like to be able to find out which people link various sets of friends.

– But I view some of the more corporate online tools as not well suited for my needs (I don’t need stats, and am more interested in the cross-links / cross-posts between people).

– Might quite like some fancy visualisation tools (more interested in the people connections, rather than the brand or top person).

– Quite likely to try a trial version.

– What if you’re interested in the network of people, or finding out who the key influencers are?

– Any “user graph functionality” “A told B told C” out there?

– Existing online solutions targeted at brands only capture high level / statistical info – do not care about fine-grained networks of referrals?

Level of IT competence and equipment:

– Just me and a laptop.

Accredited social network users (Well known people):

– I would like to keep track of my bad fans.

– Especially those that post negative, libellous or abusive content.

– I might want to know if they post that content to any other well known people’s accounts.

– Having identified them, I might want a history of everything they post ….

– …. even if that material is later deleted from online servers.

– I would like to keep track of my good fans.

– Find out who the top posters are.

– And who else they are fans of.

– And what else they post.

– And which posts they like the most.

– And who else they are talking to.

Level of IT competence and equipment:

– Possibly one or more server machines, if they are keen and of the IT type.

– Might need to architect a solution with several import machines communicating to one DB, or alternatively, importers etc communicating with each other via some streaming system.

Accredited Social Network Users (Brands / Small companies).

– Find out who the top consumers are.

– And what other brands they are fans of.

– And what else they post.

– And which posts they like the most.

– And who else they are talking to / recommending (Mentions in feeds etc).

Probably already covered by online tools.

Level of IT competence and equipment:

– Possibly one or more networked server machines, if they are keen and of the IT type.

Might need to architect a solution with several import machines communicating to one DB, or alternatively, importers etc communicating with each other via some streaming system.

– Probably not in an initial product release.

Other people who wish to monitor online activity quietly.

– I need to keep an eye on people, and comprehensive database of their posts.

– But I either do not want to, or am not allowed to delegate that to any other third party or company.

– I might be more technically knowledgeable, and simply want something like a database / spreadsheet import / export tool, rather than a fancy GUI.

– I might be able to maintain my own databases.

– I might not want to run a logged in app, but leave something sucking up data in the background.

Level of IT competence and equipment:

– I can see the hackers wanting to throw several machines at this.

– Multiple clients to a common DB possible.

– Customised networking layer also possible.

– Probably not in an initial product release.

Identification of product differentiation / selling points.

– Many of these functions already performed by online apps like HootSuite (etc).

– Which perform API amalgamation & forwarding,

– And have written tools to translate between the various API’s.

– And have their own apps which run

– But which have to maintain their own server farms.

– And do not provide any of the underlying database info to the end user.

– Although they often export stats.

– Their rate limiting is probably a closely guarded secret.

– There is no way that I am going to win against the existing companies who have already written online apps.

– However, many of those tools are corporate rather than individual based, and brand oriented.

– They have much more computing clout, and deal with statistics, rather than individual feeds, let alone individual links of connections between them.

– Some of them do maintain copies of individual social media feeds

– Although the rate limiting, and how they maintain the copies is probably a closely guarded secret.

– It might be possible to use some of those online apps as upstream amalgamation feeds for my app,

– My app currently concerns itself with maintaining a large database of existing posts, and users thereof, and deals mainly with the connection between users, rather than any of the rich media, or tracking large scale stats.

Distributed Database possibilities.

– If many copies of my app could communicate, then potentially, it might be possible to maintain a distributed copy of a reasonable chunk of a social media companies database.

– Problem with this is, customers need to co-operate, and won’t be interested in the same products / people.

– Starts to look a bit like Tor. I have no idea what the legal implications are.

– Definitely possible to have multiple webby import engines feeding a common DB …

– But then this is probably not lightweight enough to run on a single laptop for the small users.

Identification of Software Development / Deployment / Licensing Model.

– Most of the existing tools are online.

– They have a subscription model.

– They don’t sell software – just the use of the software running on their servers.

– I have no desire to set up a web hosted application(s), which uses cloud based linux hosting to co-operate with other linux apps.

Develop a natively hosted Windows (later Mac / iOS / Android) App.

– Licensing and copy protection already well understood.

– There are good copy protection and install systems that are already out there.

– Initial development costs and familiarity of tools excellent (I’ll stick to what I know).

– Parts of the architecture may change if/when the product takes off.