At LikeFolio, we have a love/love relationship with our social […]
Data & Scoring Overview
SCORING
VOLUME data is a Score from 0 up. The higher the score, the more tweets were counted in that data series.
Volume is the scoring used for Consumer Mentions (Buzz) and Consumer Purchase Intent Mentions.
Sentiment data is a percent positive score from 0-100. The higher the score, the more positive tweeters are about the brand/product/cashtag.
This is the calculation (# of positive tweets)/(# positive tweets + # of negative tweets). Note that neutral tweets are not used in this calculation.
Sentiment scoring is used in Consumer Happiness Mentions. Sentiment can be delivered as this score, or as a volume score by Sentiment type (Neutral, Positive, Negative)
METHODOLOGY
Consumer Buzz (Mention Volume): A tweet is counted as a mention if it contains a reference to a brand/product owned by the company.
Note-- A single tweet that contains two brand mentions (e.g. “Should I take my iPhone or iPad? ”) would count as +1 for each of the brands, which would then combine to count +2 for the company as a whole.
Consumer Purchase Intent Volume: A tweet is counted as purchase intent and counts toward Consumer Purchase Intent Volume totals if it contains a reference to the brand/product along with a word/phrase that indicates purchase.
We maintain a database of purchase intent words/phrases for each product type. For example, “got a new iphone 7” would count as a purchase intent mention for iphone 7, and “Made reservations for our Disney World trip!” would count as a purchase intent mention for Disney.
Consumer Happiness: A tweet is counted as a positive tweet if it contains a word/phrase in our list of consumer positive words/phrases, and negative it it contains a word/phrase in our list of consumer negative words/phrases. Tweets that do not contain words/phrases from either list are counted as neutral.
Aggregation: Each tweet score is assigned to a brand/product. For volume calculations, we add up all brands/products owned by the company into a single company score.
For sentiment calculations, we do the same (all positive mentions/all positive + negative mentions) for the company
Increments: Score by company is available in the following increments: hourly, daily, weekly, monthly.
Weighting: “Base” data is converted to “Tweets per million” data by dividing the “base” data count by the adjusted total number of tweets in the LikeFolio system for that same time period, then multiplying by 1,000,000.
(adjusted total excludes large product/brands whose shifts can drastically alter the number of tweets in the system, like “iphone” during Apple events for example)
The purpose of this “tweets per million” approach is to normalize the data over time periods of varying total twitter usage. This way our data is consistent whether it is evaluating a high-volume time period like a weekday evening, a low-volume time period, or whether or not Twitter grows in usage.
Breakdowns: Score can be broken down by brand inside of a company. Score can be broken down by division of company.
Rollups: Score can be sent as total score for sector by combining companies in that sector.
Timeframe data metrics: We return the requested score (sentiment, cashtag volume, etc) as described above for the requested time period.
All data can be accessed via the LikeFolio Research Dashboard as well as the LikeFolio API. Each of these data points are used to identify trends and shifts in consumer behavior, allowing us to provide signal to investors via our Earnings Predictions and Opportunity Alerts.