3 ms·

Those things are easy to purchase in the short term, but it quickly becomes very expensive to purchase 10% week-over-week increase of Twitter followers, Faceboo

by dmor 13y ago

Those things are easy to purchase in the short term, but it quickly becomes very expensive to purchase 10% week-over-week increase of Twitter followers, Facebook likes, web traffic, etc (I know, I've tried this before). The Mattermark Score reflects long term growth trends, short term spikes aren't enough to move the needle.

minimaxir 13y ago
"Weekly scores are averaged over 90 days," for companies that haven't received press until atleast 90 days ago is hardly "long term". That's literally 12 data points per startup, way too little for any type of smoothing to have an accurate effect.
- dmor 13y ago
  We collect 6 key data points on each startup every single day for the score, and track over a dozen more that we may factor into it in the future. This is just the beginning.
awwstn 13y ago
I think that could make sense in the long term, but with YC companies on demo day, I'm not sure it does. Most of these companies appear to have <1000 followers on the services you're measuring. If Likes and Followers cost a dollar each, then it wouldn't be too expensive to buy a spot at the top of your list in this case. It's interesting that the company at the bottom of your list (Reebee) has 3x the Twitter followers and 6x the FB followers that the company at the top of the list (OneMonthRails) has. If Reebee started out (i.e., when you started tracking them) with a couple hundred followers on each of these services and OneMonthRails started out with none, then Reebee had the deck stacked against them, needing to add 100x+ the number of followers added by companies that began with no traction, during a limited time frame.
- dmor 13y ago
  We're very aware of these issues and our algorithm accounts for them. You're welcome to sign up for free trial to poke around and see why each company's score it what it is. We provide graphs and details on all the data points for each company. http://www.mattermark.com/app/signup http://www.mattermark.com/app/signup

by dmor 13y ago

minimaxir 13y ago

"Weekly scores are averaged over 90 days," for companies that haven't received press until atleast 90 days ago is hardly "long term". That's literally 12 data points per startup, way too little for any type of smoothing to have an accurate effect.

dmor 13y ago
We collect 6 key data points on each startup every single day for the score, and track over a dozen more that we may factor into it in the future. This is just the beginning.

awwstn 13y ago

I think that could make sense in the long term, but with YC companies on demo day, I'm not sure it does. Most of these companies appear to have <1000 followers on the services you're measuring. If Likes and Followers cost a dollar each, then it wouldn't be too expensive to buy a spot at the top of your list in this case. It's interesting that the company at the bottom of your list (Reebee) has 3x the Twitter followers and 6x the FB followers that the company at the top of the list (OneMonthRails) has. If Reebee started out (i.e., when you started tracking them) with a couple hundred followers on each of these services and OneMonthRails started out with none, then Reebee had the deck stacked against them, needing to add 100x+ the number of followers added by companies that began with no traction, during a limited time frame.

dmor 13y ago
We're very aware of these issues and our algorithm accounts for them. You're welcome to sign up for free trial to poke around and see why each company's score it what it is. We provide graphs and details on all the data points for each company. http://www.mattermark.com/app/signup http://www.mattermark.com/app/signup