Andrew Chen › Comments

"Cool concept, and lots of room to growh. My guess right now is that it's buyer-side constrained and there aren't enough companies buying crowdflower services to keep up with the massive number of time-rich people online who can do menial tasks in exchange for discounted prices on virtual crotchless fursuits." - Andrew Chen

"Devon, there is no advertising allowed on my blog!!! :-)" - Andrew Chen

"Fair enough, and I agree that TC loves to sensationalize stuff :-)" - Andrew Chen

"There's definitely a bunch of crap on Google, but they have done a great job over the years of prioritizing user experience over monetization. In particular, they display less ads on searches, they have a mysterious "quality score," and they even check your capitalization when you're creating ads." - Andrew Chen

"have you used offers before? If not, try them, because you'll learn something: They are awful. I am both defending the offers industry in that I think there is some value there, but at the same time, offers are terrible. I think the link you sent me is full of denials, and doesn't take responsibility for any part of the obvious user experience problems." - Andrew Chen

"I don't think any of the recent actions Facebook has been taking will pose an existential threat to the offer providers in the short-term... it's the medium-term where continued degradation of the UX will cause FB to take big, unilateral steps..." - Andrew Chen

"That's right, current offers don't make sense at all. Right now the industry is just opportunistic, and not thinking about the long term." - Andrew Chen

"Matt, I think that if you guys don't get squeeky clean soon, Facebook will regulate you guys out of existence because of the public uproar it's causing... You guys need to get a LOT better on the user experience sooner rather than later. They'll do to leadgen what they are doing with bad actor apps and advertisers. I bet you guys have 6 months to figure it out, max." - Andrew Chen

"Yes, and similarly I've been sitting in some user research sessions with normal people and they love social gaming. The point of my article is, I think it's all here to stay. Maybe some user cleanup will be required to make all the players more accountable, but social gaming is a fundamentally sound trend and we'll see multiple IPO/M in the space." - Andrew Chen

"Thanks for the comment James. Just read your section of Adam's book "Viral Loop" btw, fun stuff. I think you are probably logged in via Facebook for Disqus and there's something weird there. Might want to try the Twitter OAuth instead." - Andrew Chen

"Yep, I totally agree with you on the user experience issues. My point is just that I think it's possible for the offers industry to actually provide end user value, though clearly they are far from that, and I want to give the rationale for how something like that might happen. Thanks for the comment :-)" - Andrew Chen

"Yes, and I think there are ways to start a company that MIGHT turn into a VC-backable company and have that optionality all the way. That's definitely nice. But I think that makes things extra hard, because not only do you need all the weird VC constraints, but you also need the constraint of getting profitable earlier rather than later. It's just hard (but not impossible)." - Andrew Chen

"I wrote some more notes in my reply to James Hong in the above comments, but my point is that there is a rather defined "pipeline" of what it means to start a venture backed company in Silicon Valley. First off, you kinda have to be here in the bay area. Then you have to get connected to the right angels, develop the right story (and tech and company), and then learn how to deal with VCs. Then you need to be introduced to enough VCs, have the right pitch, and get the right coaching (and raw results) to get it done. It's really hard, and kind of an obscure set of skills. There's lots of steps, and it's not clear to me that any amount of "practice" in adjacent models really substitutes for just doing it. And I'm not trying to make a judgement for why one way is better than the other - they are just different models, but the VC-backable model happens to be obscure enough for the entrepreneurs that would like to be doing it, they should just go for it and not set their sights on something..." - Andrew Chen

"Yep - I agree with all of your points and don't think they contradict any of mine. The bulk of my post is targeted at people who would rather be shooting the moon, but set their sites lower for training wheel purposes :-) Another nuance that didn't come through in the article is the idea of VC-backed versus VC-backable. I think whether or not you take venture money, putting yourself in high-growth, competitive markets is a recipe that's different than doing something targeted at profitability as early as possible. And you learn different lessons, and yes, the skills may not be transferable either way. The worst case scenario, as you mention, is a person who takes VC money and then regrets it. There is a lot to regret in that situation :-)" - Andrew Chen

"really viral apps can have high DAUs because they are acquiring lots of users, on their way up the S-curve. So viral apps with crappy retention can have high DAU/MAU ratios even though they are churning through a lot of users, without an indication of retention." - Andrew Chen

"This same problem happened for me. Here was the fix, for anyone who is pulling out their hair! Hi, I solved the "biosinfo" as well at the missing file at the copy on the virtual disk. 2 solutions : Do-it-by-yourself and modify usb_prep8.cmd. Line 519 replace : xcopy %xpsource%!xpdir!\!btfile! %tmpdrive%\%btdir% /i /k /y /h with: xcopy %xpsource%\%xpdir%\!btfile! %tmpdrive%\%btdir% /i /k /y /h Line 521, replace : xcopy %xpsource%!xpdir!\!btfile! %tmpdrive%\%btdir% /i /k /y /h | tee.bat -a usb_prep.log with: xcopy %xpsource%\%xpdir%\!btfile! %tmpdrive%\%btdir% /i /k /y /h | tee.bat -a usb_prep.log Line 525, replace : copy /y %xpsource%!xpdir!\%%H %tmpdrive%\%btdir%\%%I with: copy /y %xpsource%\%xpdir%\%%H %tmpdrive%\%btdir%\%%I Line 527, rmplace : copy /y %xpsource%!xpdir!\%%H %tmpdrive%\%btdir%\%%I | tee.bat -a usb_prep.log with: copy /y %xpsource%\%xpdir%\%%H %tmpdrive%\%btdir%\%%I | tee.bat -a usb_prep.log" - Andrew Chen

"I think you need hundreds to enter the funnel or more, but it depends on the level of differences between a test candidate and its alternatives (I'll leave the math to someone else, but you can read about this discussion in "landing page optimization" books/posts). So usually that means dozens of users per day so that experiments close in a few days. One advantage of early optimizations is that the differences are often quite high, and thus you can use smaller datasets to show differences. When the differences are smaller, and/or the response rates are lower, then you need more data run over more days." - Andrew Chen

"I think doing a pivot - when, how - is one of the hardest choices to make in here. This discussion probably deserves a full blog post, but I find there's often a point in optimizing products where you're stuck and can't find anything beyond small improvements - then it's often time to try a new market position or value proposition rather than shifting pages and buttons around." - Andrew Chen

"I think Gambit's user base is skewed, but is probably representative of offer platforms in general imho. But good point ;-)" - Andrew Chen

"I think Gambit's user base is skewed, but is probably representative of offer platforms in general imho. But good point ;-)" - Andrew Chen

"There are about 300 ad networks, and people know it's a massive $500B industry (offline+online). So there's a lot of very smart PhDs and geeks chasing it, probably thousands of talented people total - I'd argue the ad network industry is far more efficient than Netflix, no matter how smart they are :-)" - Andrew Chen

"There are about 300 ad networks, and people know it's a massive $500B industry (offline+online). So there's a lot of very smart PhDs and geeks chasing it, probably thousands of talented people total - I'd argue the ad network industry is far more efficient than Netflix, no matter how smart they are :-)" - Andrew Chen

"There are about 300 ad networks, and people know it's a massive $500B industry (offline+online). So there's a lot of very smart PhDs and geeks chasing it, probably thousands of talented people total - I'd argue the ad network industry is far more efficient than Netflix, no matter how smart they are :-)" - Andrew Chen

"One potential implication is that social gaming is mostly made of 13-30yos anyway, and while the ARPU numbers are high for the 30+ set, it won't counteract the full MASS of young users ;-)" - Andrew Chen

"average revenue per PAYING user per month" - Andrew Chen

"Hi Cristian, is Flairbuilder your project? Would you email me when the new versoin is out, so that I can review it? Thanks." - Andrew Chen

"In reality, the design will probably be both low-resolution, low-fidelity, and very inaccurate ;-)" - Andrew Chen

"Ha! I hadn't seen the Napkin Look and Feel toolkit. The same should be made for Adobe Flex and Javascript/jQuery, since those are often incorporated into web prototyping these days." - Andrew Chen

"Thanks for sharing! Very interesting to dissect presenting to end customers versus execs, and how you might want to change things. Absolutely they have very different motivations. The same might be said for investors versus press/analysts versus team members. All very different degrees of what you want to emphasize." - Andrew Chen

"Laura, I agree that of course, you should increase the fidelity as you go up. But early on, imho, the most important thing to get feedback on is the core concept, some of the high-level interactions and flows, in what order things happen, etc. At the point where you are testing interactions on one page, then you have to get down to a much more granular level. This is especially true if you are doing things that may be strongly affected by "polish" elements like button color, headlines, help text, etc." - Andrew Chen

"Wow I have never heard of an Excel-based prototype. Care to share one? I would really enjoy seeing it." - Andrew Chen

"Wow I have never heard of an Excel-based prototype. Care to share one? I would really enjoy seeing it." - Andrew Chen