There are two kinds of people with pet causes I appreciate
The altruists, who dedicate years of research and learning to finding and giving freely medical or other breakthroughs to make a better world. I am glad they exist, please give them my tax dollars, but I stopped trying to fix the world around 35 years old.
The other kind is patio11 kind - sorta altruistic - but getting paid altruistic. Since my Damascean conversion at the age of 35, I appreciate this kind too - you cannot have my tax dollars but you can persuade me to buy
The world needs both kinds. Well, it needs the first kind, it will never be short of the self-interest altruism. And since his pet cause is about making me money, I like his self interest.
That would refer to the Apostle Paul's conversion on the road to Damascus. Paul was a murderer of followers of Jesus, but then became a follower himself. So, I assume the commentator believed in altruism through commitments to poverty by the altruistic, big government spending, or something, but now believes in a more competitive free-market form of it.
I studied theology and scripture under a pastor and a rabbi for two years and they never described Paul's conversion as "Damascean". The contextualization is poor as well.
Obscurity and speaking out of your head for your own benefit rather than to be understood is postmodern frippery not the mark of a solid education.
Man, seriously, chill out. Neologisms are fine. Just like "patio11ian rate increase" is fine.
As far as I can see you're just engaging in personal attacks and character assassination here while not saying much in the way of argument besides, "I'm more educated than you."
I mean, you could just say something non-inflammatory like, "I'm confused, what do you mean by 'Damascean'?" and then when given an answer reply, "Oh I see, thanks. In theology circles 'Damascean' isn't really used to describe that situation."
Anyway here's a counter-argument: the Bible has long been a work of popular culture outside of religion, this isn't a theology circle or a church, and if people want to lightly abuse the stories and language in it then so what?
(I would have responded to your post where you say "strawmanning", but HN wouldn't let me. EDIT: You've now amended that post to say that the word is actually "Damascene". If you knew this, why all the drama? Something tells me that this was learned recently via Google.)
When I read your other posts they came across as mean, arrogant, and holier-than-thou. This could be a limitation of the written word and my emotional projections onto it, it could be an accurate perception and a lack of self-awareness on your part, or it could be a mixture of both. My vote is for both. I'm willing to believe you thought and felt you were writing in jest and in mirth and were not trying to be cruel, but I'm saying it didn't come off that way to me and what seems like a few others.
I would go through and highlight each sentence that reads as mean or belittling in some way, but it's literally every sentence, usually multiple times. (Except here, where you're being polite but defensive, which makes sense since I did criticize your actions.)
At the very least, please just try to understand that telling someone that their words are "not the mark of a solid education" is abusive.
which will far better argue and illustrate myriad points on the invention of language, its beauty and wonder and value to differentiate us as a species and to bring us joy and fulfillment and illustrate the challenges and quirks that make its study such a joy and illustrate language's evolution as a thing of wonder.
Then again, you see something like the Rossetti archive
and you are awe struck at the meticulous rigor and academic sensitivity it takes to understand permanence and context and format and physicality as intrinsic to the right of artistic expression.
Would these works have as much meaning if we OCR'd them, slapped them in times new roman, and beamed them to your twitching sweaty little palm sized cubes? Probably not.
It took a lot of time, a lot of arguing, a lot of hard questions and personal growth before I felt qualified to make statements like, "I don't think its a good idea to build a game engine that uses a Barthes-esque narrative decomposition to dynamically construct compelling storylines for video games." I don't think a religious education qualifies you as arbiter of language. I think his word is fun and quirky. Interesting. Highly relevant and timely.
You're not responding to anything I said, but rather you're strawmanning me to make a point you've been waiting for an excuse to make.
You might also consider using the return key to break up your text a little better. You'd be less likely to be perceived as a colleague of the unabomber. (Edit: they fixed it)
You don't need to be an arbiter of language to find out what is and isn't a word. Damascean isn't a word. The word is Damascene. The phrase is Damascene conversion. Using the unword Damascean only harms the form and function of the prose.
This term applies to both kinds of people. The first kind of people you listed are also self-interested.
The only difference is what kinds of things motivate them. The first kind are motivated by the pursuit of knowledge and/or sharing it with others. The second kind have a similar motivation, but include money into the mix.
What's up with ABsoluteMaybe? Are you still using it. I implemented my own EF storage interface, but after six months the queries started bogging down and I had to kill it. I've been meaning to get it back running.
Your out of the box storage is just a serialized file, right? The whole data set is stored in memory? How does that scale?
Out of the box is just a flat file, yes, which absolutely won't scale over time. I haven't been using the project recently (was out of the .Net world for a bit) but may be using it again soon and with a need for actual production scale. I'd love to chat more about the limits you hit. Ping me at [email protected].
If you're looking for ways to make things sound condescending, you'll reliably find them. Even in comments that say "you're awesome, don't forget to raise your rates".
It's condescending for an authority on digital marketing consulting to think your portfolio and talent is enough warrant charging 10 to 100 times more? I think it came off more respectful than arrogant.
It's arrogant and condescending to assume one knows more than, and offer unsolicited advice to, someone who just helped build the platform that raised over a quarter billion dollars in 6 months, when one has achieved nothing of the sort.
OP built an e-commerce platform that raised $1B in a matter of months. Top comment sells bingo card software to school teachers. Who is the authority again?
The wrong lesson from this would be to immediately try and get your sites' performance to similar numbers. But what isn't explicitly stated in this post is that at the root of it, these guys first figured out which metrics matter for their goals. Only then did the optimization pay back. In their case, a few seconds can make decent impact because of volume. Someone with 100 uniques/day may have a better metric than load time to optimize.
There are always a thousand things that you can prematurely optimize (page load time, HTML, conversions, etc). The trick is figuring out what the highest ROI thing is for you to optimize right now given your specific point in time. If you get 100 visits/day set metrics around sign up, retention, or whatever is going to build you up to 1K/day. Then figure out how to get to 10K/day, and so on.
I see a lot of concern for integration with API, hosting, etc. but what I don't see is disgust. Disgust because to become president of the US, you must participate in this monumental financial circus of who spends what. Is this what people want? To know the best candidate will need to raise over a billion dollars to compete? Campaign costs should be in line with needs. Remember when Color got $40 mil for doing basically what other startups manage to do with a couple thousand dollars? Why would HN readers champion this kind of spending? The reasons are probably in line with the SuperPAC reasoning, which is "in order to change the game, you must play within the game's rules". Sure. But when do you say enough is enough? In Canada, our candidates spend on their campaign, but dollar-for-dollar, they are getting more done with less money. Somehow.
Or you could view this as a fascinating case study regardless of all that. I'm personally horrified that Obama got re-elected but this post is cool as hell and I'm absolutely fascinated by this work and appreciate that Kyle took a few hours to compose all of this and share it.
Kyle and the team who worked on the campaign are awesome. And if I could have voted for Obama, I would have. My criticism isn't to do with any specific person. The campaign spending just seems out of line with other countries.
Not to defend the current campaign status quo, but when comparing the US to Canada it is worth pointing out the population of the US is ten times that of Canada.
True, but my problem is that democracy is supposed to be about everyone, and when you start eliminating people based on their propensity to generate millions of dollars you eliminate a lot of potentially great leaders.
If a national candidate cannot be organized and inspirational enough to generate donations, how can they hope to succeed as president? I would argue that fundraising is much easier than, say, passing a major piece of legislation.
The US 2012 election is expected to have cost over $6 billion when all is said and done.[1] Canada's election cost $45 million.[2] The difference is 90% less per captia. Americans spend a lot more proportionally.
If you think anyone on the campaign was being paid that handsomely you should look at our FEC reports. I took a 15% pay cut to work twice as many hours and so did all of my colleagues. The money went to where it was most needed and the proof of that is that we won the election.
My point is that a large chunk of that billion dollars went to buy TV ads, which in turn helps pay the large salaries of TV newspeople who question where the money goes.
The Obama campaign spent less, got more air time, and did it through leveraging technology to optimize their ad buys. Honestly, it should have gone into the salaries of people like Kyle Rush and the hard working dedicated technology workers for more artisinal keyboards.
It's the Circle of Life
And it moves us all
Through despair and hope
Through faith and love
Till we find our place
On the path unwinding
In the Circle
The Circle of Life
A solution is to try to find ways to change the game.
An independent-reasoning educated populace is one way to fortify against manipulation.
Reply written in the NYC Public Library reading room. This inscription is out front:
"The City of New York has erected this building to be maintained forever as a free library for the use of the people. On the diffusion of education among the people rest the preservation and perpetuation of our free institutions."
Although Canada has been blessed with a twenty-year streak of good federal governance, I wouldn't say Canadian politicians get more done with their campaign spending.
I would say that their spending is strictly capped at the national and local level by Elections Canada, which means the major political parties simply aren't allowed to get into an advertising arms race. Without the caps - who knows?
Practical, but that sort of tight restriction on free speech is a very, very hard sell down here.
For me, being a citizen and being able to become president under fair rules is important. The whole idea of democracy is that you get to elect a leader and that leader could even be you. You might be a good leader - the best even, but it won't mean a whole lot of you can't generate 1 billion in campaign contributions. Something about that doesn't seem right. Plutocracy is more appropriate - and I don't mean that in a sarcastic way. Sincerely. How is it democratic that a smart, progressive, informed person can't run against someone less informed?
The scarce resource in being a successful politican isn't really money, and it certainly isn't being smart, progressive, or informed. It's the ability to persuade people. Regardless, I think the idea that anyone could be president is kind of sweet but not very relevant. Most people can't or don't want to be elected to president of their high school class, let alone anything bigger. Democracy simply means that you get to vote, not that anyone could be president.
This misses a key point, which is that the Obama campaign relied heavily on small donations from _individuals_ in contrast with the heavily superpac funded Romney campaign which pulled from a small set of extremely wealthy donors (admittedly conjecture based on being an avid political news junkie). So, we're talking about crowd-sourcing a president; this is an encouraging direction, even if one finds the overall amount spent to be distressing.
Obama had the same superpac corporate support that Romney did. The whole emails asking people to donate $3-$5 was to bring his average donation rate down.
It seems to me that disappointment in the current state of campaign finance reform and interest in how a campaign dealt with the current reality are not mutually exclusive.
At the risk of sounding like a troll, honest question here. Why wasn't the CVV part of this page? Seems like a lot of headaches and bad press could have been avoiding by just actually including it (and you know, less fraudulent donations).
Great work by the way with the optimizations. As patio11 noted, add a 0 or two to your rate and point people to this blog post.
Checking CVV isn't required, but may result in a higher transaction fee for the merchant. The campaign likely made a choice of "fewer fields = better conversion".
It's 100% certain that it would be less conversion. It's also nearly certain that they were accepting illegal foreign donations. In this case, conversion probably shouldn't be the deciding factor... (Note I'm sure the OP had 0 say in this, just wondering if there is an explanation other than better conversions)
This isn't just about Obama, it's about all of the campaigns. Obama was just the most visible and highest raising, with a higher likelihood of attracting foreign money. It's also not just the CVV, but many many things that aren't being done (that aren't required yet but should be).
> It's also nearly certain that they were accepting illegal foreign donations.
A CVV wouldn't make a difference for that, and given that US citizens live all over the world, there's realistically no way to prevent illegal foreign donations.
What would you use? Name? American have all sorts of names. Credit card billing address? Americans live abroad. IP? Americans live abroad. SSN? Fakeable (ask a restaurant dishwasher in NYC), and I dunno if you can ask for it.
It is a tiny minority of people that live overseas. It's reasonable to make them jump through an extra hoop or two to keep the rest of the money out of our elections. Note this isn't just Obama's campaign, many others have the same problems. The only thing partisan about this is that Obama is the most visible and also the most likely person to be receiving these illegal donations. As such, he should have been working harder to prevent them, not being the most lax in the group.
The GAI report recommends election officials:
-Integrate safeguards to limit the solicitation of money from foreigners by requiring donors with foreign IP addresses to provide proof of U.S. citizenship before they can proceed to the donate page
-Immediately require campaigns to use industry-standard anti-fraud security technologies including, but not limited to, the Card Verification Value (CVV) and a rigorous Address Verification System (AVS)
-Immediately require all campaigns to retain and disclose identifying information on all online campaign contributions, including those falling under the $200 nondisclosure threshold currently allowed under federal law
-Address the threat of "Robo-Donations": The absence of industry-standard anti-fraud credit card security features render campaigns more vulnerable to so-called “robo-donations.” Robo-donations are large numbers of small, automated donations made through the Internet to evade FEC reporting requirements.
> Though not required by law, OFA requires a copy of a valid passport from any contributor who has been affirmed as eligible but donates with a mailing address outside the U.S. If they do not offer in one in a timely manner, the donation is returned.
The big question, as with voter fraud, is really "is this happening much at all?" I suspect the answer is similar to that of voter fraud - "nope".
On the other hand, any US company can make massive donations to 527s, superpacs etc., and it doesn't matter a whit where the shareholders or executive team are based.
What I heard (sorry no source) was that the Obama campaign was doing analysis post-donation to make sure they were valid donations. This seems like a more realistic approach given the scope of the problem.
Little off topic here, but did "project dreamcatcher" have as much of an impact as you guys thought it would? I didn't hear much about that toward the end of the election. What kinds of grassroots stories were generated from it? I didn't really get the sense that there was as cohesive a narrative this time around. With hurricane Sandy everything leading up to the election seemed rushed and confusing.
When running numbers like the A/B conversion improvements, how do you model or isolate external factors like "conversion rates increase as election draws closer". Is it correct to assume that no one is willing to continue to run original, unoptimized pages for some percent of visitors as a baseline?
Within any one particular A/B test, the changes in the true propensity to convert over time don't matter, since you're randomly apportioning folks into the alternatives, such that the only difference in the samples is whether they're exposed to A or B. Over the course of an A/B testing regimen, this doesn't apply, and you should be skeptical of results reported like "CR went from 10% to 15%", but a string of winning A/B tests is virtually statistically certain to be driving true wins and clients love hearing the change-over-time thing, so even if you're getting a tailwind from the environment one tends to report the numbers anyhow.
It is a poor allocation of resources for most clients to continue to run original, unoptimized pages just to be able to satisfy one's curiosity as to what portion of an improvement observed over, say, 6 ~ 24 months is due to one's own effort and what portion is extrinsic to the business. There's both a direct financial cost to maintaining old code branches and a regret cost of not giving them the best you're capable of serving. Round numbers: to add one sentence of clarification to this blog post, Team Obama would have had to give up ~$5 million in donations.
I suppose it is a fair thing to say that when listing the sins of marketing, reporting accurate increases in metrics that one cannot be sure they are responsible for would appear rather low on the page.
Can you describe how you set Akamai to host the site? I'm familiar with using S3 to host a static site, and using a CDN to host assets, but it's not clear to me how a CDN can be set to host a full static site. Did contribute.barackobama.com just mirror something like contribute-s3.barackobama.com?
Our Devops team configured Akamai so they would have specifics for you (@scottvdp).
With most CDNs you have what's called a pull zone that points to your origin server on a different ___domain. So when the user requests something like cdndomain.com/whateverpage the CDN looks for the file at pullzonedomain.com/whateverpage and then it serves what it gets back from the pull zone.
Akamai has more advanced configuration options than Amazon Cloudfront. One example of this is that our payment processors did not actually live on contribute.barackobama.com, only the static files did. So we used Akamai to setup a reverse proxy to POST the donations to a different ___domain. Without this reverse proxy it wouldn't have been possible because the same origin policy forbids cross ___domain, client-side POST requests.
In regards to Akamai and not Cloudfront - Akamai had a different offering than cloudfront. JdotP mentioned in this thread that we used Akamai GTM, it allowed us to do a lot of fun failover management which cloudfront didn't offer. It also allowed for SSL.
Would love to know more about the reasoning and discussion behind going from a 1 page process to 4 steps. (ie: The typical purchasing mindset vs. donation mind set)
Four steps is better for all sorts of purchases. Each bit of informations entered is a bit of effort invested in the process, making each additional step more difficult to quit (sunk costs). Making the first steps easier (and, especially, putting the payment info at the end) means the user is eased into the process. People also like seeing progress, hence the 1-2-3-4 thing (though that can also be daunting; seeing that there are many more steps to go: test this). Of course, while I think this is solid general advice, test/whatever for your own situation.
That's exactly what we did with voteforchange in 08. Scott Thomas and Chris Hughes were the brilliant team behind that. Then they went back to a one page form, like the only thing good about voteforchange was that you could make it cutesy and brand it really easily for lots of different demographic groups and not the shit tons of UI research and optimization that went into creating the shortest possible path for the registrant, collecting partial data, saving application state, etc.
My personal experience with local politics was that when the chips are down in a local race people don't see the value of technology to organize. When you're working on a local level you have to think like a local businessman/woman. Politics at the local level has traditionally leveraged free work. Volunteer work. Above all. Get the sap. When the sap proves useful, use connections to get them a position up the food chain instead of paying for it. Save money. Save money. Don't spend. You're going to be a one term shill. You won't make it even if you make it. You'll be more in debt than you can imagine anyway. Sell out to those with money. Don't pay for anything. Don't pay for anything. Don't pay for anything. Use promises. Lie Lie Lie. Don't pay. Get it free. They live next door to lexis nexus and they don't "get it". Its wasted. That's what's down the pipe in 2014 and that's when your shit is gonna go down. That's when you'll need someone to fix it. You know who's gonna be on the phone? A volunteer programmer who's taking 1 class this semester because he failed his fuck... no fuck that. You know who's gonna be on the phone? Liz lemon. With no staff. There IS no fucking saturday night live. Liz has been pretending there is this fucking awesome comedy show and there isn't. There IS. NO. FUCKING. TECHNOLOGY. at the local level. Its not worth doing because if the president pays fuck all then we're not going to bust our budget getting, "data" pfff. Throw more bodies at it!
If I was worth the skins and meat and seed I came into town with and I didn't spend half of it getting wasted with the local loser and I didn't brush and pop and settle well. I might have to run for president instead of ending my career framing what I should have... fuck it. Petraeus should be given . a role. He's kind of a good leader. Even if he can't entomologyze with the rest of US.
I enjoy sitting around feeling useless. That's why I choose to work in politics at the highest levels. Thank you so much for helping me achieve my dream of one day getting hired to sit around with my thumb up my ass for months at a time while you have endless fucking meetings. This is my thank you. Because I know how much you love offending people and nonsense. Thank you.
Well bend my foot on the way through door. I suppose I aught to have sympathy. I will pluck every silver haired thread from any dollhouse I'm given and I'll find the last stomped mouse on my way to work and I'll eat any shithead that thinks his men are more qualified to pre-suppose a pho-castrophe. I find men who think they're qualified to lead because they're matched with rough women. well. I find it amusing. I am rough women.
Know this. I will be somewhere in the world doing cool shit when the DNC shits in SVP's EC2 bucket. I won't know exactly when his heart dies, but I'll feel it. Like thousands of voices screaming out and being silenced. I might be in the bathroom taking a piss, looking in the mirror as the camera slowly pans in on my face and I will breath in slowly, and out like darth vader, but I'll know.
That's what I hate about rails. You spend so much time filling out YAML and reading documentation and making sure you've got the right gems for the job and that you didn't accidentally go down a rabbit hole that by the time you're ready to build something it looks like the same bucket of sand everyone else is waiting for the tide to come and collect.
If it's anything like the non-profit space, the multipage checkout, if done right, can allow you to psychologically hook the donor in and yield higher completion rates. (Clicking a donation option is much lower commitment than typing your name and address as a starting point.) It also allows you vary the checkout steps based on what the donor selects. You can then treat a $10 differently than a $500 donor to optimize conversion rates for each cohort.
I've approached this problem myself for an ecommerce platform. In certain use-cases, less pages/steps doesn't always mean more conversions.
For this instance, the large photo background of Obama in action evokes more emotion than the formal portrait. Also, it's a very simple and clear call to action. In this case, the donation amounts are their add to cart.
It's also very important to understand how users got to the donation page. From the article, it would appear that most of the "sell" was done via other media (social, email, traditional links). After their A/B tests, I could assume that they are leaving the sell of "why" up to the users. This allows the designers to focus on conversions.
Lastly, the four steps aren't traditional "checkout" steps. Each step focuses on very specific information, eliminating clutter and keeping the user's attention. Too many elements on one page can be a bad thing.
Cheers! After looking at it more (I don't have experience donating to political parties), I realized (I may be wrong) that alot of information are actually mandatory and affect your taxes(?) down the road. Was confused why all necessary info required, like addresses and employment information. (When I was think of 1 page, I was think just donation amount + CC # and CVV)
But I look forward to reading any numbers of before and after you share!
Address is pretty standard for anyone accepting payments.
As for employer and occupation, we are legally required by the Federal Election Commission to do our best effort to gather the donor's employer and occupation. Doesn't have anything to do with taxes.
Ahh I see. I know some processor now forgo the requirement of addresses, eliminating alot of fields and barriers even more. It may take quite a while to get used (and feels scary too).
The rationale I've heard before is that people are put off my large forms because they see the amount of information they're going to have to input.
If you split it up it looks like less, and by the time the user realises that step 3 has 10 fields they've already progressed through two pages and feel "invested".
I can't speak for them, but we do know that big forms are intimidating and tend to induce the feeling of, "Ugh, I don't have time for this now."
Even though there are technically more steps, they are unambiguous, digestible and feature a clear sense of progression. User usually don't mind multiple steps as long as they make sense within the context of the goal at hand.
In the first UI the users are overwhelmed with all the information they have to fill out in order to make the donation. In the final UI the process has 4 steps.
Thanks for posting this fantastic write-up! I'm looking forward to reading more.
Does anyone know if other team members from the campaign have publicly discussed how other aspects of the online campaign worked? (Such as analytics, ads, CRM, social media, etc)
First of all, congratulations and thank you for doing the work you did to re-elect the president. Your write-up there is really interesting. I have a couple questions:
1. According to FEC reports, Blue State Digital made millions of dollars working with OFA... With that in mind, why is it that the campaign tech team had to build the redundant donation API? You would think for millions of dollars BSD probably should handle that, no? I only ask because in all the post-election stories a big theme is the "we decided to do it in house"...which makes sense...but begs the question - If you were "doing it all in house" then why was this third party vendor making so much bank?
2. Since we have you here on HN, can you tell us (very generally) what you and the other OFA Chicago Digital/Tech/Analytics people are going to be doing now? Going back into politics? Forming political/non-political startups?
I do not agree to it. Has any one seen the Obama campaign spending on software and infrastructure. They spend half a million dollar on Microsoft software (the highest spending), and then all the highlights of the campaign are coming from Open source software.
What was the benefit of spending on Microsoft software. Why it is always that money is spent on Microsoft and benefits are highlighted of using open source software.
Kyle, great post. You said the final donation was made via an XHR POST, did you guys have a plan B for users without JavaScript or were the numbers small enough to ignore?
Yes. It was very simple. We just put the URL to the vendor hosted platform's POST endpoint in the donation form's action attribute so that if you didn't have JavaScript the form would POST like a normal form.
Reading this write-up after having read Nate Silver's post[1] today on how the Republican party has a harder time getting Silicon Valley talent leads to an interesting thought. Going forward the party that can successfully target the younger, smarter base will have a SUBSTANTIAL advantage in developing better systems. Assuming there is not some ceiling on how much better technology can impact efficiency, these systems will translate to exponentially more effective platforms for the party with the brightest youth.
You could also argue that Republicans will slowly peel off disaffected or disenfranchised talent as technology in politics becomes less of a cottage industry for seasonal fruit pickers and they have the lawyers to push against the true ceiling you're talking about... the legal hurdles and the stagnation of innovation that comes with winning and scaling.
Seems a little dubious to claim that the increase in donations was a direct result of increased speed. The nature of the election cycle seems like more donations would take place as campaigning ramped up and the election date drew nearer. Were those kind of external factors accounted for in your analysis?
The increase in performance that I mentioned is directly due to the performance increase. We built identical pages on both platforms and tested them against each other at the same moment in time.
The only difference between the two pages was the page load time. Because we removed all other variables (time, appearance, functionality, etc.) we can say that the 60% performance increase lifted conversions by 14%.
I have a question about the donation API though, I feel like I am missing something:
We settled on a very simple solution of turning the hosted platform into a REST API. The only big change was adding JSON as an output option instead of HTML... We consumed the newly created donation API using JavaScript on static HTML pages which were served by our CDN (Akamai).
What operations did this API handle? Were donation submissions being sent from the user's browser to this API?
By the way, I think there might be a bug with your PJAX code on kylerush.net - when you click links to other parts of the site or to read other articles, the page title displayed in the browser stays as the initial title from the initial HTTP request.
The API was quite simple. Essentially it just processed a donation. It had some other functions to it that involved saving a users payment information to their account as well.
Yes, the donation submission was an POST request executed with JavaScript (AJAX) from the user's browser.
Thanks for noticing the bug! I'll have to check that out.
I've some questions to further understand the optimization process.
- In general, how did you measure improvements in conversion? What metric were you using? Something like LTV or the % of conversions each variation got?
- How exactly did you verify that the new version had a 14% larger conversion rate? Did you run an A/B test with both?
Kyle, amazing post! You mentioned that the 60% speed increase led to a 14% donation increase. You used webpagetest.org to measure the 60% speed increase. How did you measure the donation increase? Is there some sort of A/B tool for CDNs to help quantify the impact of speed improvements on CR?
We used Optimizely (https://www.optimizely.com/) to conduct most of our a/b testing. Optimizely works on the client level (browser), not the server level (CDN).
Optimizely is the best a/b testing software I've ever used. It's flexible and easy to use. I suggest you check it out.
How did you implement the fail-over to a different amazon region when one fails? I'm interested in common approaches to this SPOF and what have been done in this case.
*seeing the discussion that is going on around this post, it almost feels off-topic to post a technical question :)
I am curious where this will lead the next election, I think it is quite possible that 2billion mark will be broken as everyone is looking at this an will be getting ready for the next round in 4 years. It is quite depressing actually
Wow, all of that with Jekyll as the CMS generator. I've been using Jekyll to generate static sites and have been proud that mass traffic rushes don't bring them down...but I haven't used it for anything that generates real revenue
Very nice write up! :) It's really nice to see A/B testing promoted, as well as, a case for static sites integrated with web services. Those are the types of technologies I'm really looking forward to seeing grow in our industry.
Too bad that there were no checks in place to determine who is actually contributing. For that reason, many non-citizens were able to contribute, some from questionable places, and even Osama ibn Laden had posthumously donated, receiving congratulation later from Barrack's wife thanking him for his donation [1].
Since Romney was rejecting contributions that could not be confirmed as of where they came from, this race in terms of fund-rising was skewed from a get-go. Imagine a marathon where everyone runs on their feet and you -- against the rules -- are using a bicycle. Guess who's gonna win?
The IP address and name are irrelevant. Should US citizens living in Pakistan be disallowed from making campaign contributions because they have arab sounding names?
The only thing that matters is whether the donations come from a US credit card. As far as I can tell all of these farcical contributions came from US credit cards.
Nothing in those articles reliably supports (i.e. beyond a couple individual anecdotes) the contention that a) the Obama campaign had "no checks" or b) that Romney was rejecting unconfirmable contributions.
"Walker said he used his actual street address in England but entered Arkansas as his state with the Schenectady, NY, ZIP code of 12345."
In fairness, that's fraud on Walker's part not on the Obama campaign's part. There's a limit to what kind of checks you can feasibly do. By the sounds of it, requiring a billing address in the United States seems reasonable.
> There's a limit to what kind of checks you can feasibly do.
Actually, this one would've been easy to catch - we know which states have which zip codes. The company I work for checks this in web forms, but I don't find that such validation is common.
> By the sounds of it, requiring a billing address in the United States seems reasonable.
Problem is, it isn't. US citizens living abroad have every right to donate.
Actually, I think that zip code is valid :) according to a quick Google search at least.
But I admit I hadn't thought about citizens living abroad; that makes it more understandable there would be these issues. It's certainly a case of (perhaps) incompetence over malice anyway.
Edit: my mistake, I missed that he claimed the zip was in Arkansas.
You would need some sort of photo of their passport to verify citizenship for a foreign national. You would have to securely store thousands of these images for reporting purposes. It would be a legal and technological impediment and extra cost involving a large amount of effort for very little payoff. This was easier to limit with the online store. They just didn't provide international shipping options.
Why does it not surprise me that someone who coded a newsletter for Drudge Report - a conservative cesspit of a website - is sounding like such a sore loser? Hmm.
(All the websites you linked have a similar conservative slant. I mean, the Washington Post? Seriously? Some people have no shame.)
> Too bad that there were no checks in place to determine who is actually contributing.
Why is this bad, exactly? Political leaders in the US undertake actions and policies that can affect the entire world, esp. in regards to military action, so why wouldn't some people abroad want to donate to a campaign? Also, given that there are several million non-citizen residents of the US, they deserve just as much input into the political process of the society they are a part of.
Well, the legality of a thing doesn't determine whether or not it is desirable. Why do you feel that non citizens donating to a campaign is bad outside of legality?
