Posts from Web/Tech

One of my favorite features on Twitter is the ability to retweet something with added context.

Like this:

This is a terrible idea https://t.co/5yHt65M8qY

— Fred Wilson (@fredwilson) December 22, 2017

I do this a lot:

“NYC Female Founders Office Hours” https://t.co/KrwxJV6m0H

— Fred Wilson (@fredwilson) December 7, 2017

There is a lot to be thankful for this year https://t.co/bPFSkGjPdX

— Fred Wilson (@fredwilson) November 23, 2017

“This was to prove they belong, that they really can play with anybody.” https://t.co/s1pg9Tn8dq

— Fred Wilson (@fredwilson) November 23, 2017

You get my point.

There are a lot things Twitter can do to make Twitter better but getting rid of the Quote Retweet is not one of them.

I’ve got mixed feelings about Twitter’s experiment with allowing longer tweets (280 characters vs 140).

Like many users expressed on Twitter (of course) yesterday, I’m quite fond of the 140 character limit.

I don’t like the constraint when I compose tweets, but I love it when I consume them.

There are few things that make Twitter unique, defensible, and essential (contrary to many Twitter haters, it is all of those things).

At or near the top of the list is the sort bursty stream of information Twitter presents to the consumer.

There is no other place where I can consume a firehose of information across so many topics as quickly as I can on Twitter.

Just looking at these tweets from Jack and Biz, I am not sure 280 characters is going to be a good thing for the consumption experience.

This is a small change, but a big move for us. 140 was an arbitrary choice based on the 160 character SMS limit. Proud of how thoughtful the team has been in solving a real problem people have when trying to tweet. And at the same time maintaining our brevity, speed, and essence! https://t.co/TuHj51MsTu

— jack (@jack) September 26, 2017

Originally, our constraint was 160 (limit of a text) minus username. But we noticed @biz got 1 more than @jack. For fairness, we chose 140. Now texts are unlimited. Also, we realize that 140 isn’t fair—there are differences between languages. We’re testing the limits. Hello 280!

— Biz Stone (@biz) September 26, 2017

On the other hand, I think running experiments like this is the right thing for Twitter and every app out there to do.

And they can’t run an experiment like this without telling the world about it. I’m seeing longer tweets in my timeline. They can’t keep that a secret.

The one piece of advice I would give the Twitter product team (who explained themselves in this blog post) is that they should test 160, 180, 200, 220, 240, 260, and 280.

I suspect they will get the biggest impact with slightly longer tweets but not all the way to 280.

I frequently run out of characters in my tweets. But generally not by a lot. If I had another 20 or 40 characters, that would reduce my character limit frustration significantly.

It’s also easier to introduce gradual change to a user experience than radical change.

And doubling the tweet size is a pretty radical change.

So I’m glad everything is on the table at Twitter in an effort to improve the user experience. That’s how it should be. But I’d be careful about this experiment and test a wider range of tweet sizes if I were them.

A number of our portfolio companies that have free web/mobile apps that are monetized by advertising have offered a low-priced subscription offering that removes the ads and, often, offers offline sync on the mobile apps.

Here are some examples:

• Quizlet Go
• Duolingo Plus
• SoundCloud Go

I am sure there are other examples in our portfolio but those are the ones I am most familiar with.

I like this model a lot. As Duolingo said in their Duolingo Plus launch communications, it allows a free service to remain free for those who can’t pay for it.

It also allows those who don’t want the ads to remove them. And other value added features, like offline sync, make the subscription offering compelling for power users.

Pandora’s Plus offering is a good example of this approach and, because it is a public company, we can take a look at the numbers.

In Q4 2016, Pandora had roughly 4.4mm paying subscribers out of roughly 80mm total users, only about 5% of its user base.

But if all of those 4.4mm subscribers are to the low priced ($5) plan, then they would generate $265mm on an annualized basis. I assume that the subs revenue number is a bit larger because there is also a $10/month plan. So let’s say subs revenue is $300mm.

Pandora’s total revenue is about $1.4bn a year so subs represents over 20% of the revenue even though only 5% of the users take the subs offering.

So if you have a free ad supported service with a lot of regular power users, you should really consider adding a low priced subscription offering.

It will diversify your revenue mix and gives your users the ability to opt out of the ads if they want to.

Ten years ago yesterday, August 23, 2007, Chris Messina suggested that using the # sign in a tweet could be used to group tweets about a thing (his example was barcamp).

Here is the tweet in which he made that suggestion.

how do you feel about using # (pound) for groups. As in #barcamp [msg]?

— ⌗ChrisMessina (@chrismessina) August 23, 2007

There is so much I like about this story.

1. A user suggestion became a feature simply by other users adopting it.
2. Twitter took it a step further by recognizing the hashtag as a “thing” and making it hyperlinked in the service.
3. The hashtag feature took off as Twitter users thought it was cool and started using it actively.
4. Users showed that hashtags could be used for many more things than groups.
5. The hashtag moved beyond Twitter into most popular social media applications (ie Instagram) and now has become the standard way for users to “tag” something.
6. The hashtag has moved beyond software into our language and media.

I was an investor and board member in Twitter at the time and the Company was tiny, not more than twenty employees and probably a fair bit less. The engineering team was maybe half of that, led by Blaine Cook. They were struggling to keep up with all of the growth of the service and yet they had the foresight that the # and @ symbols that users had adopted organically were important. They prioritized recognizing them in their software. That was a big deal.

Biz Stone, now back at Twitter, blogged about how this happened yesterday and it is a short and good read.

The Twitter story is a complicated one, with lots of ups and downs, but I believe it has been and remains one of the most important services ever built on the Internet and hashtags is one of the many reasons that is so.

We are fortunate here at AVC. We have mostly civil and respectful conversations. People behave themselves here. That is sadly not the case everywhere.

I don’t know what the people who post comments like this are feeling and thinking. It is horrible. Awful. Hateful. Hurtful. Painful. Disgusting. Disturbing. And a lot more.

If you operate a large social media service like Twitter, Facebook, or Disqus, you get to see stuff like this every day, hundreds of times a day. It is a view of humanity that is deeply upsetting.

Disqus, which is a USV portfolio company, where I serve on the Board, and which operates the comment service here at AVC and at millions of other websites around the globe, has been working on scaleable solutions to this problem.

They posted an update yesterday on what they are doing to combat this problem.

Here are some excerpts from that post:

The Disqus Platform supports a diversity of websites and discussions; with such a large network of publishers and commenters, having a policy against hateful, toxic content is critical. While we do periodically remove toxic communities that consistently violate our Terms and Policies, we know that this alone is not a solution to toxicity. Oftentimes these communities simply shift to another platform. Ultimately, this does not result in higher quality discussions, and it does not stop the hate. In order to have a real, lasting impact, we need to make improvements to our product. Which is why, if at all possible, we work with publishers to encourage discourse (even unpopular or controversial discourse!) while helping to eliminate toxic language, harassment, and hate.

Over the past several months, many passionate folks have reached out to us about severe violations of our Terms of Service. With the help of our community, we’ve been able to review and enforce our policy on dozens of sites.

We appreciate all of the help and feedback we’ve received and we are excited to continue to partner productively with users and organizations that are passionate about fighting toxic content and hate speech. To improve our efforts, we’ve built a Terms of Service Violations Submissions form. This form is a way for users to explicitly share with us when they’ve found a community to be in violation of our terms. In addition to reporting individual users (which helps moderators know who in their community is perhaps exhibiting toxic behavior), you can now report directly to us when you think there’s a publisher/site we should take a look at. When we are made aware of potential violations, we review them internally and make a decision about whether or not to allow the site to remain on our platform.

This isn’t a small scale matter; we know that to have a meaningful impact across our network, we need to build solutions into the product. With that in mind, we’re committed to building tools to make the moderation experience easier and better for publishers (and commenters, too).

Here are some things that we’re working on:

• More powerful moderation features. We’re working on two features right now, Shadow banning and Timeouts, that will give publishers more options for managing their communities. Shadow banning lets moderators ban users discreetly by making a troublesome user’s comments only visible to that user. Timeouts give moderators the ability to warn and temporarily ban a user who is exhibiting toxic behavior.

• Toxic content detection through machine learning. We are working on a feature to help publishers identify hate speech and other toxic content and then handle this more effectively.

• Commenting policy recommendations. While we already provide suggestions for how to create community guidelines, we’ve realized that we can be more proactive and more assistive to our publishers. We’re working on helping our publishers expose their custom commenting and community guidelines by making them more visible to their readers and commenters.

• Advertiser tools: Just like publishers do not want toxic content on their sites, we know that advertisers do not want their content to display next to toxic comments. Leveraging our moderation technology, we will provide more protection for advertisers, giving them more control over where they display their content.

If you think this is a simple problem to solve, you are sadly wrong. And if you think that Disqus and USV and I don’t care about solving this problem, you are wrong about that too.

In reaction to the Trump Administration selling us all out to the big telcos, I am dedicating this blog post to proxy services, which we should all start using.

A proxy server is a computer (in the cloud) that redirects your Internet browsing and other traffic away from your ISP to a specialized service that can do a lot of things for you.

Here is how you configure Chrome to proxy your traffic to something other than your ISP.

Here is how you configure OSX to proxy your traffic.

And you will need to select a proxy service. There are plenty of free ones out there. But you get what you pay for. If you want to have a proxy that is reliable and won’t sell your data, you should consider paying for a proxy service.

Here is a good roundup of some popular proxy services.

I am sure that many AVC regulars are using one or more proxy services and I encourage all of you to share with us your favorites.

My partner Andy and my friend David are doing an “interview cast” which is “a 24 minute conversation about one topic, with one expert.”

They call it The Riff and their first episode is about the evolving news media landscape with CNN journalist Laurie Segall.

Here it is:

I posted the discussion my partner Andy and I did at the Upfront Summit last week.

There were other great conversations at the Upfront Summit.

This discussion with Mark Cuban was great. I totally agree with Mark that we need more tech companies to go public and have been saying that publicly for several years.

This was a fun twitter exchange yesterday:

@fredwilson agree, but how do they better monetize that influence and impact?

— Scott M (@SouthofAmerica) February 9, 2017

I can’t help but think that competing with Facebook and Google in the online advertising market is an uphill battle for Twitter. They have done a great job at building a $2bn annual advertising business that pays the bills and generates positive cash flow. I know the people who have built this ad business and they are world class.

But given that Twitter’s strength is influence and impact, not page views and clicks, is there a business model that compliments the ad business that Twitter should be leaning into?

As always on fun fridays, the action will be in the comments. So let’s get this discussion going.