In testing the Dataswarm Analytic Engine on stock analysis and prediction, we are tracking the events leading up to the IPOs of a number of “Unicorns” – Uber, Lyft, AirBnb and Slack – to see if there is anyhing that can be learned from the social (and other) media coverage to predict what the IPO may go in terms of pricing and subsequent performance of the stock. We covered the Lyft IPO over here.
At any rate, for Lyft there were mixed emotions immediately pre IPO, and even though they raised the IPO price we were not seeing the level of hype and enthusiasm we had expected for an IPO. We thus suspected the stock price would not have such a great time immediately post IPO and that proved to be the case, as the stock, after a short “pop”, dropped to near IPO price levels and went under IPO price 3 days after the IPO. But we didn’t really have anything in the database to compare it to apart from our exeperience , as the dataswarm system has never been used for this beforehand.
Uber was different, as we now had a pattern – Lyft – in the same sector, to compare it to. So it was interesting to see how it was doing on the same relative day in its IPO process. Tracking was far more informative this time as we had Lyft’s statistics in so many areas to compare it with. The graph above shows quite an informative summary view, this is cumulative relative sentiment over the monthe before and straight after the IPO, both are started at 100%.
As you can see, the curves are remarkably similar in shape overall, but the details are interestng. In detail the Uber IPO started off stronger (Lyft had a bigger lift later in its process) but they got to the same point a few weeks before the IPO and seemed on roughly teh same track – but then Uber sentiment started to drop off rapidly compared to Lyft. There are a number of potential causes and we have some views of the major drivers from our data now, but what was clear is they were – in our estimation – probably not going to have as good an IPO as Lyft, and indeed they had to lower the projected IPO price. Again, if the Lyft data was predictive, we expected the stock to drop post IPO, and the “pop” to be fairly underwhelming – which proved to be the case, and it did even drop to below IPO value by the end of Day 1 of trading.
Post IPO, Uber sentiment has continued in the doldrums whereas Lft sentiment again lifted, and the stock price behaviour for both in their post IPO Week 1 looks similar. In Lyft’s case, at the end of Week 1 it started to drift lower for several weeks, so if this pattern is in any way predictive so will Uber’s for a few weeks. We will see….
It will be interesting to see if the patterns for Airbnb and Slack are similar, or if Lyft and Uber were “sector specific” in some way.