Weekly Traction Tracker: Who Is Hottest Among 1,100 Startups? | Hacker News Reader

Weekly Traction Tracker: Who Is Hottest Among 1,100 Startups?

daniellemorrill.com

25 pointsdmor12 comments

It's a surprise to see my dead startup TutorialTab as #6 in weekly growth! Seems like your algorithm is sensitive to fluctuations in sites with really low Alexa rank (TT is at 13,890,005 right now). FYI!

HA! Okay, I'm looking into that right now, will get back to you.

Update: It went from 21,369,458 to 13,900,665 in the past week - so I probably need some sort of lower bound for these. In some cases that can indicate an up-and-comer, and I am even using a log scale for the delta so this a pretty serious outlier. Thanks for letting me know it is dead though, i will remove from the DB.

You could try only taking account companies that are ranked under 100k (like Alexa does on their own website).

Many of the companies on that list have ranks somewhere in millions, which can fluctuate quite easily, and many of the companies have been around for months or years.

Edit: Accelgolf seems to be dead too. Blog & website broken and no updates for two years on twitter, linkedin indicates the founders have moved on.

Cool! As someone who was pretty critical of the FRC post from yesterday I want to make sure I spoke out to praise this. This kind of data, rather than an index, is much more interesting to me.

A few other jumbled thoughts:

* How are you tracking these? I assume it must be pretty messy collecting this data and keeping it up to date.

* Could you do something with Google Trends as well?

* Clearly some of these (whereberry/swish/etc) are highly correlated across channels, implying some sort of news event. Would be really interesting to see the ones that only gain in one without the others, and what could cause that. This kind of turns into Google Finance for startups.

* Agreed the LinkedIn data is intriguing, if only because it has a bias towards much more established companies -- maybe because it's focused on job-seekers?

* Lastly, one philosophical point: If you go by the mantra of "what gets measured gets managed," does this end up reinforcing the silicon valley hype cycle and encourage startup CEOs to get "hot"?

Thanks, I'm collecting it in a massive database and I love hearing all the ideas of what could be done with the data.

As to what gets measured gets managed - legitimately increasing traffic, audience, brand awareness etc. is all about the top of the funnel. Considering how many startups I meet with who are hemming and hawing about doing marketing, only to later end up dead, I think this could be a very good thing. We'll see.

Several in Twitter Following % Growth have <200 followers and a few with only 10s of followers. At these numbers 100%+ growth may not necessarily be an indicator of "hottest".

That's true. I messed with other ways of weighting it, or just putting the absolute number of new followers (e.g. Dropbox gained 70,000) but already has 2.3M. Thoughts?

What would be interesting is establishing benchmarks from the historical data you have collected around Likes/Follows/etc. Then using the benchmark, compare the weekly or monthly change for each startup relative to its previous Like/Follow count

This is interesting, but the problem is this is all a certain type of startup. Whisper, for instance, is "hot" both in funding and user growth, but it's a demographic that wouldn't trigger these indicators - kids at 19 don't have linked in, and they aren't linking their FB or Twitter accounts due to the type of app it is.

Basically this is a Klout for startups? BTW, Klout API might be useful in figuring out the social traction http://klout.com/s/developers/v2

I would think that website traffic growth is the least important of these metrics, definitely not the one to lead the post...some of these are web services, others look like native apps...it's not an apples to apples metric at all.

Twitter/Facebook activity would be a more useful measure...not just followers, but how active the accounts are and how often they're handling customer inquiries. And yes, this can be gamed by buying followers...so what? Include the deltas and let readers decide the legitimacy of a gain of 50,000 twitter followers in a week for a no name startup

Thanks, I have that activity data and will be packaging it up for a future post.

Reply on news.ycombinator.com