MerrowAnalytics.com - Sell Project

Do you really know your Twitter audience? Merrow does. Merrow Analytics provides FULLY AUTOMATED deep analytical reporting on your, or any Twitter account for just five bucks a report. Customers sign in with Twitter, then by using Stripe.com make a one time report; Or they may choose to setup a recurring daily, weekly or monthly report for long term tracking of high value Twitter accounts. All clients have to do is type in a target username, choose the report type - and at almost no cost to them, they have an actionable report to present at the next meeting. Merrow is heavily targeting the SEO market as an initial vertical, with branding focussing on the 19-27 year old social marketing crowd - attempting to disrupt the social insights market by offering an option that won't break the bank, but still delivers professional looking presentable reports that lead to noticeable ROI gains. Merrow is built on Node.JS, Express, Redis and Twitter Bootstrap, and is primed for desktop, mobile and tablet display in any web browser. Pages are generated using simple {{replacement}} operators, and the complete source code is fully documented in well formed English sentences. Merrow needs a loving team to see her through to a proper full launch, once it's running it requires virtually zero maintenance (other than DB backups), and is already setup to be scalable-scalable (potentially 100-million+ users). Current MVP reporting status: 1. Timeline Deep Hashtag Analysis (Beta 1.2.2) By aggregating carefully weighted hashtag data from not only a timeline, but also from it's 0->n deeply connected timelines, we are able to produce an actionable ratio of retweet, favorite and new follower conversions per hashtag in the local Twitter neighbourhood. 2. Timeline Heat/Mass Analysis (Beta 1.0.2) Tweets are the atomic object in question; each one carries a 'heat' value which is calculated from it's recent traffic generated, weighted by the importance of that traffic (ie. Bill Gates retweeting is more important than your grandmother retweeting). Heat cools over real time (hours, days, etc.), and forms into mass. A tweet that consistently generates traffic will be massier than a tweet which bursted traffic for a short time, thanks to the exponential nature of heat to mass transfer in the algorithm. 3. Follower Clustering (Alpha 0.0.4 Implementation Proven At Scale) Once an audience is divided into 1->n clusters by most similar timeline comparison, it is trivial to discern human labels for actual groups of followers rather than hypothetical groups (ie. 'people who like punk music' over 'teenagers'). Useful in determining if your social marketing strategy is actually connecting with the target audiences. By creating simple Bayesian features and using timelines as landmarks in a learning kernel, we can easily determine similarity between timelines. Current issue is rate limit requests from Twitter. 4. Outreach Recommender (Alpha 0.0.8 Implementation Proven At Scale) Each Twitter account is allowed to create X tweets and follow Y new users every day. In order to maximize the conversion return on those X tweets and Y follows, the outreach recommender provides a list of Y users who are not currently followed that are most likely to follow back or convert on tweets made. Heavily uses search, issue is rate limit requests from Twitter. 5. Will-Follow-Back Estimator (Alpha 0.0.8 Implementation Proven At Scale) Used as part of the outreach recommender to determine a coefficient between 0 and 1 representing the likelihood that a given user who is not already following, will become a follower if followed first. (Note that descriptions here are brief, and are represented here only at the highest level for easy buyer comprehension purposes)