This number shows how our data predicts the story will grow. It tells you how many more Facebook interactions or Twitter Influencer shares we expect there to be over a specific period of time.
How it works
We detect a trend based on the size of increments in interactions between data points (how fast the story is growing), weighing the most recent increments higher than past ones. This trend is then linearly extrapolated to produce an estimate of the number of Facebook and Twitter interactions a story will have earned by the end of the prediction period.
The longer we observe a story's social interactions, the further ahead we can predict with reasonable accuracy.
How Accurate is the Prediction?
The accuracy depends on how much we know about the story (i.e. the timeline of interactions) and how far into the future we want to predict. This means that our forecast is based on the age of the story - the older the story is the further we predict.
For one hour old stories, our forecasts are, on average, accurate to within +/-21% of the predicted score. For three hour old stories, our forecasts are, on average, accurate to within +/-12% of the predicted score. For stories older than 24 hours, the deviation drops to just 1%.
Many stories get most of their social interactions within the first 48 hours after being published. After that, the number of new interactions diminishes a trend that is also reflected in the predicted values.
* For all stories younger than one hour, i.e. time since publication is less than an hour, we forecast the number of interactions they will have when they are two hours old.
* For stories that are between one and three hours old, we forecast the number of interactions they will have when they are six hours old.
* For stories that are between 3 and 12 hours old, we forecast the number of interactions they will have when they are 24 hours old.
* For all stories that are older than 24 hours we forecast 24 hours ahead since the last interactions that we captured for that story