Twitter Cuts Off Employee Access to Its Metrics | Hacker News Reader

Twitter Cuts Off Employee Access to Its Metrics

recode.net

70 pointsdanso43 comments

I'm amazed that it took this long, given that the # is material to how the stock trades. (Look at their volatility on earnings day) Why give thousands of people access to information whose inadvertent release could cause public disclosure problems? (As a public company, material information needs to be released to all investors at the same time) Even an accidental mistake could cause big problems, let alone insider trading.

Seconded. The company I used to work for had stories of a gong on the sales floor that would sound when someone made a sale. Once they were publicly listed, though, the gong went away, since anyone on the floor would have an idea of sales for that month and therefore be subject to insider trading.

I immediately concluded that being in a publicly traded company is boring. I work for a school now. :)

There's tradeoffs. Public companies are more boring, but having permanent capital does bring a little more stability. Different things for different people. To borrow from Peter T, it's much easier for private firms to get from 0 to 1. Public companies seem to be better on globalizing, where standards and process are important. Neither is universally better. (I've moved to steadily smaller companies in my career)

I'm not sure what you mean about being publicly traded gets you permanent capital? THe only capital you receive is from the stocks your company sells. All the subsequent trading doesn't benefit the company - after the IPO proceeds (and subsequent offerings that may be issued), the stocks are only a liability to the issuing company.

Technical nit: stock is equity, which is absolutely NOT a liability in the technical sense of the term. Generally liabilities are something that has to be paid back on a more-or-less predefined schedule; equity is more "permanent" (but it's really a spectrum with points in between like "convertible debt" and "preferred debt" and such, which have some features of both.)

The point about "permanence" refers to the lack of any obligation by the issuer to pay it back on any schedule. That changes of ownership can happen without the company really knowing/caring is a huge feature of public equity markets, compared to say, a private partnership where, if an individual wants to exit, it might lead to business closure because the partnership needs to liquidate in order to cash out the partner, or the partner needs to find someone else to buy their partnership stake, which may be difficult as partnerships aren't as standardized as NYSE-traded equity with regard to reporting, control, and legal precedent.

tl;dr share capitalism has some really nice benefits that too often go overlooked.

It also means that the CEO is always busy talking to wall-street (or so says the CEO of the private company I work for).

Although personal stake at the founding is something that private companies (such as mine) struggle with as things get larger.

> Public companies are more boring, but having permanent capital does bring a little more stability

Privately held corporations can be very large too. Since there are fewer captains and stakeholders I would argue that at the same given size privately held corporations are more stable.

Ikea, for example, is private. Dell went private not too long ago, too.

There are tradeoffs. And the tradeoffs shift over time. Sarbanes-Oxley has (on the margin) made it more costly to be a public company, so we're seeing companies hold out over time. (One can are argue that this strengthened the system - I will leave that argument for elsewhere)

There are some absolutes that force a company's hand.

1 - If they have VC money and the investors want an exit, they have to sell it to someone. It's very rare that the current owners will have the money to do this. (If they needed $10mm to grow, will they have $100mm liquid to help someone exit?) If they sell to a PE firm, the problem just gets pushed down the road 5 years.

2 - There are limits to the # of shareholders a private company can have. If you believe in employee ownership, this becomes an issue as private companies grow.

Re (1), the big story of late is the growing size of private equity funds and companies' ability to raise ever-growing amounts of money without going public. Digital Sky Technologies (DST) did some early investments in facebook that arguably started the trend; Uber just raised two billion+ dollar rounds of funding without an IPO. This was completely unheard-of 20 years ago. [1]

Re (2) I don't think you're right, I think it's that there's a cap on the number of shareholders until disclosure becomes mandatory. Can you give a citation for your view?

[1] https://gigaom.com/2012/05/08/why-facebook-and-silicon-valle...

The test is "materiality": in case of information, will someone's knowledge of something (how many gong strikes took place) cause them to make a different investment decision about the company?

10 x $10 million gong hits per year for a $100mil business? Maybe that's material, but that's a matter of professional judgment for an accountant. It's not black and white.

I bet someone in legal or finance hated that gong and used going public to silence the gong. The thought that gong ringing could be used to gauge revenue is pretty laughable.

Full disclosure: I've worked in gong ringing sales departments.

Bullshit. The market would work more efficiently if this info was made public in real time. That companies keep important data secret until earnings' big PR effort is a huge market distortion as evidenced by the big corrections that occur during the big reveal. The only benefit that this distortion provides to the company is to allow insiders more opportunities for cooking the numbers and insider trading.

The market would work more efficiently, but it violates securities law for it to be selectively released. If you allow 1000+ employees to see the #, you might as well have it appear real-time on the investor relations page. (I'm ok with this)

Are you suggesting these should be open to the marketplace? That alternative puts all publicly traded companies at a disadvantage to privately held businesses.

You've just highlighted the tradeoff. The question for public money is "How much information to actually release?"

This works in a world where "the need to have Internet access" isn't considered "unfair". We'll get there, but it'll take a while.

Nonsense. HFT is already a demonstration of the abuse of high speed access to pricing data. The markets are not fair, and most of the participants in them prefer it that way.

In all fairness there is real market distortions and problems happening now based on getting data seconds prior to other sources and acting on it.

I don't understand - can't one calculate MAU from firehose data? If so then all companies who bought that data can access this elusive MAU metric.

MAU includes users who passively consume content and do not post content. Deriving MAU from tweet data alone will vastly underestimate it.

You can likely calibrate past firehose data with past MAU data and then you can use firehose as a predictor. I am sure there are strong correlations between the two.

In addition to this, the firehose license almost certainly includes 96 pt font printed in blood-on-asbestos specifically enumerating "publishing aggregate statistics" as one of the things you're not allowed to do with the data, on pain of a breach of contract suit with damages measured in percentage points of Google's market cap.

So why publish it, when you can just calculate and trade on the info? Though the costs of getting all the data might hurt the profitability of this plan. (The resellers charge per tweet, so you'd need to fall back to a sampling scheme, though that might be his enough.)

An 'active' user doesn't have to be tweeting, so they won't all show up in the firehose.

MAU figures are pretty dubious anyway - see the other recode story about how 4 million 'active' users vanished because iOS no longer automatically fetches tweets:

http://recode.net/2015/02/05/how-twitter-lost-4-million-user...

So, in that sense, going by unique tweeters is an even better metric. Also, since Twitter offers limited (1% or 5%) of the firehouse to anyone via the streaming API, wouldn't that be a fairly decent way to estimate things?

The data can still be correlated. Given a history of known actives and corresponding tweets, the average tweets-per-active can be determined. That may change over time but it's unlikely to do so in a capricious fashion without good cause. Say, onboarding a large class of users for whom Twitter and/or Tweets are of secondary interest. Not that, say, Google might have done that with Google+ or anything.

Just because you cannot _directly_ see the number you're interested in doesn't mean you cannot infer it by other means.

True, if only stock prices were based solely on quantitative metrics. But it's also based on human emotions and expectations.

I don't understand how/why a publicly traded company like Twitter should be allowed to hide such an important piece of data from the public and from shareholders?

It's less to do with hiding data from the public and more along the lines of controlling the data so that no one in the company gets it before the public (and thus allowing said person inside the company to trade company stock on insider information).

Isn't this solved with trading windows, which you only open up right after earnings calls? Or am I missing something?

Those don't apply to the general public. The problem is insider gets info -> calls his brother-in-law -> brother-in-law trades on early info. Illegal, but lucrative.

If you're classified as an "insider" then you can't even trade in the trading windows without a 10b5-1 plan.

So at TWTR (I'm a former employee) knowing things like the MAU would probably make you an insider, which means you can only sell on a plan.

Not knowing the MAU is a blessing in this case.

As mathattack said, public companies are required to disclose this type of information to the public all at once.

If I work at Twitter and tell you that our MAU dropped 5% this week, and earnings is in two weeks, you could short the stock and make a lot of money. It's to prevent this type of scenario. The data is private until it's released to everyone simultaneously.

I'm sure they don't. The issue isn't that they're hiding it, the issue is that every employee had access to it before the public did. From now on, most employees will learn what the MAUs are the same time the public does.

Does Facebook have similar policies? What's the level of transparency there?

That sort of information is readily available to Facebook employees as well as widely distributed. But for example engineering levels are secret.

What does "engineering levels" mean?

When someone gets promoted at Facebook, that fact is generally not shared. There isn't a big list of "who just got promoted".

I was more interested in knowing the level of transparency on a day-to-day basis on metrics such as usercount, revenue, etc. you know, stuff that could get you in a bind for insider trading.

User count - yes; revenue - I honestly don't know. There are quiet periods for restricted stock units for when you might know more than the market. When then quarterly numbers are out the blackout period ends and you can trade your stock. It's been like that in every publicly traded US company I've ever worked at.

FB have published their daily actives as recently as September, 2014.

I understand the reason they did it, though I think they should have done this way earlier. This seems to send a negative signal that you're not growing as fast and want to keep your cards close.

Reply on news.ycombinator.com