Making sense of the Uber IPO catastrophe

Making sense of the Uber IPO catastrophe

We knew this IPO was not like any since Alibaba in 2014, but that turned out to be a serious understatement.  Apart from the large offering size, the analogy with Alibaba’s generally successful IPO falls apart immediately.

Alibaba’s share price, first 24 months from IPO ($68 per share)

Nor was this a somewhat obviously mis-priced and mis-handled IPO like Lyft’s.  We have to go back to Facebook in 2012 for a disaster this big.  And even Facebook’s shares opened up on the first day, waiting until day two before breaking below the IPO price (where it stayed for well over a year).

Facebook’s share price, first 24 months from IPO ($38 per share)

First, why was Uber’s IPO unlike Lyft’s?

Uber was sensible in many ways it seems Lyft was not:


  • Uber’s initial price range was significantly below the price talk of up to $120 billion of value, and even below Uber’s last private-market share price of $48.77.

  • Uber didn’t raise the filing range or increase the deal size during its roadshow.

  • Uber’s underwriters made unusually diligent efforts to scrub the book and place shares with sticky investors.

  • Uber priced its deal at $45 per share, just above the bottom of its price range of $44 to $50 per share.

  • And this is critical – Uber’s shares never traded above the IPO price, so IPO investors inclined to flip were never in the money (much more on this below).


If Uber didn’t recreate Lyft’s deal dynamics – greedy on price and reckless on allocations – what happened?


Uber had a Triton Score of 6.35, placing it in the 3rd quintile of tech IPOs Triton has Scored over time (vs Lyft’s 5.47 Triton Score putting it in the 5th quintile).  An average Triton Score should absolutely not break price at the open, especially one so big and important as this.  As the chart below shows, even bottom quintile tech IPOs produce, on average, a 17% pop on their day-one open, and a 21% gain by the close that day.

Return from IPO price, tech IPOs Scored by Triton, by quintile (2014-2018)

Uber’s day-one open at $42 against an IPO price of $45 is a major outlier, but not unprecedented.  But the only situation even somewhat comparable among recent 3rd quartile tech IPOs was GreenSky, which had several deal- and company-specific issues that make that adverse outcome much less mysterious.  And even GreenSky finished up on day one.

Tech IPOs Scored by Triton that broke price at the day‐one open

So the question remains – what happened?

There are 5 things to consider:

1) Disclosure – some of Uber’s adverse issues were far more fundamental than the inside-baseball deal dynamics.

One way Uber was like Lyft was in its woeful disclosure, which we at Triton went out on a limb to highlight far beyond the investors we interact with on a day-to-day basis.


At Triton we have turbo model tools designed specifically to look at companies like Uber, but as we can see so clearly in the rearview mirror, not all the participants who set the prices of these IPOs have such tools, and to them Uber’s loss-making machine is a black box. 


So even though Triton’s subscribers were able to see that there are model cases to show Uber nicely profitable in the out years, there’s no reason for this analysis to be broadly “in the price” for IPO investors.


And, as you can see below, despite the rosy Bull case, the middle and Bear cases are daunting.

Triton model cases for Uber at IPO (May 2019)

In the absence of conviction on fundamentals, many investors sat on the sidelines for this IPO, which is one reason it priced so close to the bottom of its range even before it had a chance to fall apart at the open.  Separate from the macro issues of the trade war, the generally adverse tape, and overall “Uber fatigue” (which was really a thing) some reasons we heard for adopting a “wait and see” posture on this IPO were:


  • No urgency to act now, the IPO is not an investment catalyst.

  • Uber is still very early in its development and it will take years for the thesis to manifest properly.

  • There will be less risky and lower-priced entry points later.

  • This isn’t really an IPO, it’s a follow-on, as Uber is already so broadly-owned by large asset managers.

  • If you want to own this growth profile with much less risk just buy Google.

  • This emperor has no clothes.


2) Deteriorating Triton Score over time – transparency wasn’t the only problem for Uber at its IPO.  Comparing Uber’s Triton Score at IPO to its Score in February 2016 at the time of the Series G shows the Score moving the wrong way from 7.29 to 6.35.  And Uber’s below-average Scores were in very important categories.

Comparative Triton Scores: Uber at IPO and at Series G vs Lyft and Average

Also by way of comparison to the key valuation drivers for Uber at the time of its Series G, Uber missed its “Management’s Case” revenue and EBITDA projections for 2018 by a mile.

Triton “Management’s Case” model for Uber at Series G (Feb. 2016)

But Uber over-performed against “Steady State” and “Headwind” cases.

Triton model revenue cases for Uber at Series G (Feb. 2016)

Triton’s “Likely Case” model (ridesharing only) was surprisingly accurate projecting both the top and bottom lines for 2018 by assuming no improvement in take-rate over time, as opposed to Management’s Case that relied on an 8% improvement.

Triton “Likely Case” model for Uber Series G (Feb. 2016)

3) Mixed Scores blending to an average Overall Triton Score – As and Ds are not the best way to get a C average, and Uber got low grades in very important subjects. 


Some things were fine, but there was a lot not to like, so investors could see what they wanted to see and focus on the good or the bad.

Triton’s component Scores for Uber at IPO (May 2019)

4) Valuation – Uber’s shares were very seasoned at $48.77, the price set by Tiger Global and T. Rowe Price in December 2015 for the Series G (roughly $2 billion initially that grew to $8.8 billion – more than proceeds from the IPO).

Uber Series G issuance 2015 - 2018

Even though Uber grew its revenues by 4.5x over the period, and its multiple fell into a zone that compared very favorably with Lyft’s, it is notable that Uber’s share price was constant even as revenue growth slowed from 253% to 42% (see the table below).


The net effect of “adding on” to the Series G terms was to postpone the day of reckoning of a price-adjustment event that is antithetical to the current practices among the most prominent large-cap venture-backed startups. 


In the liquid public market, stock prices go up and they go down.  But in the illiquid private market where deals are very competitive and terms are negotiated over the table in that context, share price can become a constant or a one-way ratchet, not a variable that moves both ways.

Uber at IPO and Series G compared to Lyft

5) Lock-ups – back to deal dynamics and technicalities.  Why would Uber shares fall below $45 even before the open when so much stock changed hands at $45 the day before and at $48.77 over time?


Factors to consider:


  • Huge pent-up gains among early-stage investors and secondary buyers.

  • Many incentives to sell for early investors who were / are deep in the money.

  • A breakdown of civility on lock-ups in the Lyft IPO with rumors of “total return” products marketed to locked-up Lyft investors.

  • Significant new language in Uber’s S-1 amendment indicating risks related to incomplete lock-ups.

  • Traditional stock-borrow to facilitate short positions is not typically available for several days following a new issue.


Even after watching Lyft get it wrong, Uber’s underwriters couldn’t head off disaster, with three million shares offered for sale below the IPO price before the open.


The initial selling pressure wasn’t likely from owners who paid $45 for the stock the day before, the flippers never got into the money.


And it’s unlikely the huge base of owners at $48.77 would enter sell orders before the open, either.


Only when the knife was accelerating in its fall did investors re-orient to loss-avoidance to exit before it really hurt.


Where did all the stock come from that was offered before the open as low as $42?  Is the story here the incomplete lock-ups?  


No IPO like this should break price at the open.

Outlier trading against Triton Scores:

Sometimes low-quality IPOs trade against their Triton Scores for a time before correcting, like for example Jumia (5.02 Triton Score — 5th quintile) recently:

Jumia’s share price, first month from IPO ($14.50 per share)

Sometimes high-quality companies trade against their Scores the other way for extended periods, as Dropbox (7.67 Triton Score — 1st quintile) has:

Dropbox’s share price, first 14 months from IPO ($21 per share)

Dropbox Price Chart.png

What could be next for Uber?

Unfortunately, the Facebook stock price chart on the first page bears significant resemblance to the chart of IPOs with average Triton Scores for their first year of trading (flat), and especially in the first 90 days.

Return from Day 1 open, tech IPOs Scored by Triton, by quintile (2014-18)

As we can see across the broad average, gains on tech IPOs happen almost entirely inside the institutional discount (the difference between the IPO price and the first trade), except for the highest-Scoring companies in the first quintile.


In Uber’s rare case the institutional discount was negative 6.7% on day one!


If the price performance normalizes in the next two weeks and finds a level from which to regroup and improve, this episode may be remembered as just an anomaly. 


And if Uber is able to execute in any meaningful way on its larger thesis about the total transformation of the transport of humans, their food, and their stuff it’s likely that at some point Uber’s stock will reflect the magnitude of its opportunity and none of this will matter at all.

The immediate future for the tech IPO window:

When will Chewy and Douyu — both of which are filed and could have launched their roadshows on Monday the 13th and didn’t — decide to launch?


How is Fastly doing on its roadshow?  Will it actually price Thursday to trade Friday?


Where will the Slack direct listing value the company relative to recent private market benchmarks?  Will it even get done?




One longer-term possibility could be that current strategic vogue of achieving massive scale before recurring profitability will be rethought (or thrown out?) if this strategy also includes public-market IPO liquidity.