Twitter founder declares war on those bullshit statistical data on the Internet

the article is authorized by the Medium and Ev Williams curiosity journal publishing. The author Ev Williams is the founder of Twitter and Medium.

recently was quoted me, said I “don’t give a damn” sets have more users than Twitter. If you read my articles, I said I don’t care about the front there is a big “if” two characters. My words is as follows:

compared with them, Twitter’s influence is very big to the world. Between the two is at least the difference between apple and orange. Twitter is we want to become. It is a real-time information network, all things on Twitter happens in the world to see – something on Twitter discussion, dialogue between world leaders on Twitter. If so, to be honest I don’t care about more people see the beautiful on them.

of course, I could look down on the sets for many people. This software is elegant, can let a person to create and enjoy art, but also social function, this is usually a good thing. But the reason why I said those words almost nothing to do with the (also has nothing to do with Twitter). I said these words because, those who reported consumer Internet services, investment and construction of the people in the successful evaluation of site, the standard is single, it makes me feel more and more angry.

which rectangle larger?

when you ask a junior high school students, a rectangular 3 inches long, and a 2.5 inch long rectangle, which is bigger, they will tell you the problem is meaningless, unless they know the other information – that is, the rectangle’s width.

however, when we say a “big”, a company or service is basically based on the number of single – specifically the number of users over the past 30 days. Don’t even looked into what is a “use”, that’s ridiculous.

in recent sets/Twitter contrast, for example, I see Will Oremus on Slate paper haven’t abuse “bigger” the title. He explains why well things not so simple. This is an article is well worth reading, its central meaning is as follows:

then sets a bigger than Twitter? Not, it is different from Twitter. One is basically a private media and the other one is basically a public media; A picture as the center, the other is focus on ideas and concepts; They both size is large, and have been expanded.

last week is Medium hits a week – we’re about to say the biggest. The number of unique visitors from, performance is very good. And this is mainly because a highly popular widely forwarded (mainly on Facebook). But with the user on our web site, compared to the average residence time of most visitors is only for a fraction of the time, and there is almost no read something.

this is why in the industry, our first consideration is “TTR event”, namely the total reading time (total time reading). Use this indicator to measure people spent time in the article page is not perfect, but we think it can better response whether the user get the valuable information from Medium. Last week, in the case of TTR event, Medium traffic is still huge, but at the beginning of October that week TTR event data is 50% higher than last week, last week, but the number of unique visitors only 60% of the week (i.e., at the beginning of October each visitor actually read time is much longer).


Mat Honan Wired on published an article on the present situation of today’s media, very accurately pointed out that now everyone in the attention for war “in the reality. But it with the number of unique visitors to compare the performance of different media in the campaign:

according to the chart (misleading use of “reading” label), Buzzfeed traffic is greater than the New York times. But we can know from the same chart, Medium traffic last week more than the week in October. Although we only care about attention, not other important value, but that does not make too many problems. Perhaps, from the total BuzzFeed than the New York times received more attention, but we should stop to use this point to a single chart to discuss this kind of problem.

furthermore, the chart above, and you see in the media – mainly based on the comparison of site visits to ComScore, and contains only users in the United States. It’s two big problems: 1. Most of the data provided by the website operators with ComScore gap is very big. The more so the smaller the site (due to their different sampling methods); 2. Only a small proportion of total number of Internet users in the United States. Each site is now a global – only differ in the globalization degree. If there is no express only interested in that users in the United States, so just to disclose the user’s data will have a lot of misleading.

in the end, ComScore does it include application of traffic, it is not clear. I know they are within a certain range tracking application usage, but they are in the table above all is calculated by the application of the New York times (or application BuzzFeed)? (if you know please leave a message)

we focus more on people’s reading time rather than the number of visitors to Medium, because it is a world of infinite information content, in this world, you just click, you can learn numerous eye-catching things, and a succession of news to push – so someone is really take time to read, is a significant thing. Scarcity value. Although people are now willing to give the attention of the media and network overall rise sharply, mainly due to the amount of everywhere can use mobile phone network – but it is limited after all.

there is more than we agree with this view. Site called Chartbeat of Tony Haile to promote “Attention website (Attention Web)” the idea, done a lot of great work, the concept of “hits website [Click Web]” is the opposite). He hopes to measure attention, people can perform site improvement:

for high quality content publishers, they are not simply based on clicks for pricing, based on their cumulative time and attention. And this may be just what they are looking for a lifeline. In the Internet world, time is a scarce resource, we will spend more time on the good content. According to standard of the time and attention to pricing, means that compared with the site simply link bait to attract traffic, good content publishers can charge higher fees to ads on the page.

Upworthy also advocate the concept of “time on behalf of the quality” :

we would like to think so, because share let people really feel interesting and valuable content, rather than let a lot of people click on content, for we feel a sense of accomplishment. While on the basis of visitors and page views. Our performance may look better, but we are very pleased to make some sacrifices, because this is the real measure of customer satisfaction index.

and Wired the article also mentioned in the advertising industry, why more precious time:

you click is precious, your attention is very valuable, but for advertisers, the most valuable or your time, and the content of the publisher, says they want according to readers spend time instead of just the page on the site clicks for advertising. So according to this logic, users spend more time in a report, the relevant ads on the page the higher fees.

but there is also a problem , actually not really used to measure value. It measures the cost, as the value of the alternatives.

the advertiser is not the time to really want you, they want to impress you, whether consciously or unconsciously, and eventually they want, is your money).

I write this, is not the time to really want you, I am want to influence your way of thinking. If my rhetorical skills can let me finish it in less time, in order to myself and to you), that nature is better.

and Medium is not the time to really want to anyone. We are want to build a platform, let people to express themselves, here to impress others. Let people to think, to change people’s way of thinking, to teach people something, or establish an emotional connection. And these are hard to measure.

if the Medium appeared some new form of content category, or author, content is to satisfy and is very popular, but we think do not have significant value or lasting importance, this kind of situation? Such as soap or deriving of Medium. Our TTR event value may be high to Pierce the roof, may also let us feel the illusion of success, and so very happy. But take the user’s time is not our real goal, and wastes people time is run against our goal.

this is all a single standard of the problem. Just as Jonah Peretti puts it, there is no universal standard “:

I think now the industry people around, trying to find a measure of the content of universal standards. Is all about share, or is all about holding time, or is all about page clicks, or is all about access. The problem is you can only optimize one of these, so you have to choose. If you want to optimize several criteria, that you have to make a series of compromises.

in the early years when I was in the Google, remember that hangs on the wall a chart (chalk) remember the number of search. This looks like the measured value of the reasonable solution. Actually, it was emphasized to search as little as possible to take up people’s time. If your service is to provide practical value, rather than as a media, it might be a good target.

(Twitter has both practical and media attribute, and some of its advantages is its extremely concise – in addition, like Google, most of its intention is to bring people to the other place. So to optimize the use of time doesn’t make sense.

if you look at other outstanding high-tech companies, such as Apple, obviously not their goal in pursuit of the number of users and do optimization. Although network effects (and income) means that they would care about the data, but by focusing on producing the best products, they have built up the planet’s most valuable companies – in fact, one of the enterprise strategy is to produce a series of products are complementary to each other, and then put them as much as possible (with appropriate margins) sold, but also to let an user to buy for your own product.

the lessons of history and many other companies, they are admired by the number of users in, but most of them ultimately ended in fiasco.

if you can actually focus more on breadth rather than depth, most Internet companies is to be able to create better products and more value.

in the past a period is the case. In the beginning of the Internet era, people talk about the probability of page views and visitors to talk about at least as much as quantity. But before the application of rise, this has become obsolete and wasteful (” hot spot “is the same as the before that), and application of the rise of just make it more obvious.

in 2006, I wrote a blog this view was put forward.

emphasize page click quantity too stupid, but for technical reasons, emphasize the MAU/UV (Unique Visitor, visit the web site of a computer to the client for a Visitor) is harmful.

if you care about, or want to report, is the impact on the world, it becomes a very delicate. You are not in a rectangular measurement, but in a multi-dimensional space. You have to admit that things are difficult to perfect evaluation, you can only from several criteria and some case is try to learn as much as possible.

if you want to assess the value of a company, in theory could be much easier. From a financial point of view, the value of a company is its ability to long-term profitability. nullnullnullnull