Key Performance Indicators Now Available from Aletheia’s Earnings Call API
Aletheia’s Earnings Call endpoint, providing Earnings Call transcripts over a free-to-use HTTP API, can also now provide Key Performance Indicators.
Each earnings call transcript is now being analyzed by a sophisticated Natural Language Processing (NLP) engine during the processing stage. A proprietary algorithm has been developed to contextual the content in each earnings call. The algorithm searches for line items, product lines, key metrics, and more, that were stated by management as having increased, decreased, reached a point, etc. Upon finding these examples, the algorithm parses the plain remark into three components: the subject (what is being talked about), the status (what happened to the subject), and the value (the dollar figure or percentage which the subject is being measured by).
When requested via Aletheia’s EarningsCall endpoint, the Key Performance Indicator data will be included in the JSON response. The following as an example of how the KeyPerformanceIndicator property will be appended to the SpokenRemark array, listing the detected KPI’s for this particular spoken remark:
For each KPI in the response, data about the subject, status, and value will be provided in multiple formats:
- subject — the plain text of the subject that was referred to during the call.
- status — code indicating the action/what happened to the subject. Codes:
- 0 = “equals”, “is”, etc. For example, “revenue is $500,000.”
- 1 = “increased”
- 2 = “decreased”
- 3 = “Is greater than”, “surpassed”, “exceeded”, “passed”, etc.
- 4 = “Is less than”, “fell below”, “sunk below”
- 5 = “produced”, “generated”, “created”, “caused”. For example, “iPad contributed $100M in revenue.”
- value — the dollar figure, percentage, quantity, that is mentioned. If a percentage is mentioned, it will be in a percentage format (i.e. 55% would be 0.55).
- valueIsPercent — if the value is a percentage, this will be true. If not, this will be false.
- subjectOffset — start location of the subject in the spoken remark.
- subjectLength — length (number of characters) of the subject.
- statusOffset — start location of the phrase that determined the status property in the spoken remark.
- statusLength — length of the phrase that determined the status property in the spoken remark.
- valueOffset — start location of the figure (number or percentage) that determined the value property in the spoken remark.
- valueLength — length of the figure (number or percentage) that determined the value property in the spoken remark.
Using the offset and length properties above, one could easily highlight the specific KPI components in the spoken remark in a graphical user interface.
Free and Available over REST API
The service is free to use for private developers and is available over a REST API, meaning it is accessible from virtually any programming language or development stack. Register for an account to start using the service now!