Twitter Cuts Off Employee Access to Its Metrics

http://recode.net/2015/02/05/twitter-cuts-off-employee-access-to-its-metrics/

by danso • 11 years ago

70 43 11 years ago

Lur• 11 years ago

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.

minimaxir• 11 years ago

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

bhouston• 11 years ago

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.

patio11• 11 years ago

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.

joosters• 11 years ago

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...

MichaelGG2• 11 years ago

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?

dredmorbius2• 11 years ago

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.

AznHisoka• 11 years ago

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

mathattack5• 11 years ago

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.

oddevan2• 11 years ago

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. :)

mathattack5• 11 years ago

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)

eldavido5• 11 years ago

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.

jusben13692• 11 years ago

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.

guelo• 11 years ago

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.

mathattack5• 11 years ago

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)

GolfyMcG• 11 years ago

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

eldavido5• 11 years ago

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.

jusben13692• 11 years ago

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.

t0mbstone• 11 years ago

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?

oddevan2• 11 years ago

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).

mjmahone17• 11 years ago

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

pc86• 11 years ago

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.

benihana• 11 years ago

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.

kenrikm• 11 years ago

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.

badusername2• 11 years ago

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

fishnchips2• 11 years ago

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

jchendy• 11 years ago

What does "engineering levels" mean?

badusername2• 11 years ago

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.

dredmorbius2• 11 years ago

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

Twitter Cuts Off Employee Access to Its Metrics

Related Stories