Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
texasbigdata
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
77 ms
·
691.
▲
by
texasbigdata
7y ago
As someone stuck on a side project outside their area of expertise the diagram of local maximums is beautiful.
692.
▲
by
texasbigdata
7y ago
Not at the VP but at the C level you can get a feel by reading SEC proxies. Amounts, composition and sometimes methodology/justification are included.
693.
▲
by
texasbigdata
7y ago
Also "Playing to Win: How Strategy Really Works" by the old CEO of procter and gamble and the dean of the toronto business school. In addition to good strategy / bad strategy, playing to win is sort of heads and shoulder abov
694.
▲
by
texasbigdata
7y ago
Really sorry to hear that. Did you ever experience anything related (like mushrooms)?
695.
▲
by
texasbigdata
7y ago
Ok, so dumb question. Why _wouldn't_ you optimize for conversions then? Unless you want some immeasurable branding value type lift. Sorry for the dumb question but if you're implying one particular bidding mechanism benefits from
696.
▲
by
texasbigdata
7y ago
What? Assuming you live in America, "many" people's bodies are not catabolic for any meaningful portion of a given week. Not even close. You can force a catabolic state onto yourself...but how does TRT help? Note also that i
697.
▲
by
texasbigdata
7y ago
Not an expert but I've used it before. Think of it like gmail + VOIP put together. When you run an ad on google that incentizes a non-click action (i.e. call us!), to track that result, google randomly generates a phone number (I think
698.
▲
by
texasbigdata
7y ago
I've used call rail before. They don't IIRC. I believe he's using the IP (from Google?) since he says "Yet Callrail still withholds your IP address (exact location) from us.". So it seems he's taking 1) Your IP
699.
▲
by
texasbigdata
7y ago
Those metrics reveal something intrinsic though. Take cigarettes. They're bad, not liked, etc etc. The usage in 2019 reflects something deeper: their effectiveness, perhaps a cohort effect and the difficulty of quitting, and a certain
700.
▲
by
texasbigdata
7y ago
Not even remote. As a nontech worker its unbelievably helpful in a shared office too. I used to have a big screen tv and 2 keyboards and wireless mice at a coffee table in my office for just this problem, and a crazy (for normal humans) hdm
701.
▲
by
texasbigdata
7y ago
Is this cheaper than paying a recruiter fee?
702.
▲
by
texasbigdata
7y ago
Agree. While not an expert, I've tried the regression the author has done with CapitalIQ before. There's a third hidden axis which is whatever you want to define size/maturity/runway to be. That potentially explains some
703.
▲
by
texasbigdata
7y ago
Its marginally better for online shopping than Visa and Mastercard. No address and zipcode and depending on authentication settings significantly less friction to purchase. Its surprising PayPal isn't everywhere. Small data point but p