Mea Culpa on Ning
Posted by: Rob Hof on January 26, 2006
I’ve been trying hard with this blog to tap into online conversations instead of simply posting my own views, and that often means weighing in on controversial posts by other folks. But a few days ago, I made a mistake. I settled for linking to someone else’s post without fully checking out the underlying facts for myself.
In my post linking to TechCruncher Mike Arrington’s post on Ning, I said Mike’s arguments about Ning not delivering on its promise sounded reasonable. But after talking with Ning cofounder and CEO Gina Bianchini—and by Mike’s own admission in a comment on the blog of Ning lead software architect Diego Doval, whose response I’m reproducing below—it appears that two of Mike’s claims about the service were erroneous, or at least could be construed that way. (The other two might be considered a matter of opinion.)
(Update: Mike just called me and said he believes the gist of his arguments are largely correct, though he concedes that in a couple of points, he wasn’t clear enough. In particular, he thinks he wasn’t specific enough about the way in which Ning could better support common Web services APIs like Google Maps. He also would change his claim that Ning can’t help mere mortals create “unique” applications to clarify that he meant that Ning still requires some programming skill to build truly different apps. And while he still thinks Ning’s communications with users had been lacking, he says the company’s action in recent days to respond to his and others’ criticisms has been topnotch.)
Now, I did hold off posting for a few hours as I tried, unsuccessfully, to reach co-founder Marc Andreessen. And I updated the post with Ning’s response to Mike as soon as I saw it. Finally, my take was different (and still is): Even if you accepted Mike’s arguments, it was too early to write off a three-month-old service.
But none of that excuses the fact that my post gave credence to criticisms that, with a closer look, now don’t strike me as entirely justified. I just tried out the service, and lo and behold, it is pretty darn easy for a nonprogrammer like me to create at least simple Web services. (You can see my very preliminary effort here.)
As Gina herself says, Ning does have more work to do to get more regular folks to create their own online applications, especially unique ones from scratch. In fact, a major upgrade that Gina says will make that easier is coming in a couple of weeks.
Lesson (re)learned: Participating in the online conversation doesn’t excuse me from trying very hard to nail down the facts first—at least as I see them.
Update: Mike has his own update on Ning’s upcoming features, which he thinks look pretty promising.
Here's the relevant part of Diego's post on his own blog, which makes Ning's case:
The post makes four main points:
First, You have to know PHP, or at least HTML, to build anything unique on Ning.
Well, with Ning's clone feature you can get a decent level of customization without coding or knowing HTML, and we're getting easier to use all the time. However, the key word here is unique. As far as I know, you have to know PHP, or HTML, or Ruby, or something to create anything unique on the Internet, period. Ning doesn't yet allow non-coders to create unique apps with complete flexibility and any features they want, true, but neither does anyone else, and there's a simple reason for that: it's a really hard problem, and one that most definitely doesn't get solved in three months. We want to get there though, and we'll continue working to that end.
Second, there is no support for the key web service APIs out there that people are really excited about mashing up.
That's not true. Aside from support for Google and Yahoo! Maps, the Amazon API, the Flickr API, the Yahoo Search API, as well as RSS, Atom, OPML, and others, anyone can add support for whatever API they need by either uploading their own PHP files, or write their own PHP API on top of whatever REST, XMLRPC or SOAP service they want. The fact that today we only support PHP is a limitation, to be sure, but a Ruby binding is upcoming, to be followed by others such as Python.
Third, Ning keeps all of the applications under the ning.com roof. This has benefits like free hosting, but application creators don’t get control of the page to add advertising and they cannot get user registration data direclty [sic].
Also not true. Since last December, any user has been able to run an app under their own domain (domain mapping), run their own ads, make source code private, or get additional storage. In the future you'll also be able to complete remove/edit sidebar preferences to make the experience more uniquely your own.
Fourth, Ning did everything wrong in communicating with their users.
"Everything" wrong? Really? It reminds me of something I wrote back when the controversy about stealth mode happened: "Ah, those pesky generalizations." I said back then. Indeed.
Generalizations aside, there's actually a good point hiding in there-- It's true that we haven't been very aggressive in promoting our new features as we rolled them out, and perhaps we could have done a bit more "outreach" but in part we've spent a lot of time working on stability and new features.
Consider: We went 1.0 in December (the beta period lasted a bit less than three months), and removed the "invitation" requirement to be a developer -- now any user can be a developer. We have added new apps, and improved the original ones, both in speed and features. We added new services, such as domain mapping and ad runner. And more. But we didn't make a big deal out of it, mostly because we're working on a sitewide update that would address those issues in a more uniform way.