Why can't the government use the best graphic and user-experience firms to design communications. Medicare forms, for example, are awful to read. There are plenty of great design firms that can do the job. Just do it!
A/B testing is not a new concept. I find it embarrassing that our government is just now discovering how it works.
For all those "sharp-eyed" folks who are questioning a $7 price tag for a series of texts: It should be obvious that this was a research project. The subject group did not identify itself, the texts did not write themselves, and the data did not analyze themselves.
A text may cost "nearly nothing" to transmit, but transmission is the least significant portion of a text message's total cost. That includes the texts you and I send every day. Unless you have a free device and free cell service, your cost for each cellular network transaction is much higher than you presume.
A text may cost "nearly nothing" to transmit, but transmission is the least significant portion of a text message's total cost. That includes the texts you and I send every day. Unless you have a free device and free cell service, your cost for each cellular network transaction is much higher than you presume.
1
Design a better government? What a wonderful idea, and by the way, remember to brush your teeth after every meal, nudge, nudge.
The fact is that our bumbling government needs all the help it can get so nudge away and may we all profit.
The fact is that our bumbling government needs all the help it can get so nudge away and may we all profit.
2
The unfortunate truth about bureaucratic behavior is that for the average worker doing a routine, but carefully circumscribed job, the principal task is to get through each work day without drawing unwelcome attention to himself. The risk of making a mess of things may not even be on the horizon, but doing things differently, albeit more efficiently and with a better overall result, comes with all of the risk attendant to messing up, and with few, if any of the benefits. Whatever pressures for change necessarily come from the outside; but come they do, and when change comes, it's typically a top-down affair accompanied by multi-agency, multilevel task forces in which interest brokering, and responsibility and risk allocations are parceled out. At the federal level, and frequently replicated at the state level, agency officials will present carefully prepared testimony, frequently characterized as program improvements, when in reality much of their forward-looking proposals are grounded in risk management.
Once in a while, opportunities come along to make a tangible difference in how the people's business is carried out. Nevertheless, agency personnel get to see every possible permutation of what can occur, and typically more so when matters involving agency discretion increases. The essence of governance is to make choices, and that typically involves weighing and balancing competing elements within a proposed course of action, and with predictable consequences known.
Once in a while, opportunities come along to make a tangible difference in how the people's business is carried out. Nevertheless, agency personnel get to see every possible permutation of what can occur, and typically more so when matters involving agency discretion increases. The essence of governance is to make choices, and that typically involves weighing and balancing competing elements within a proposed course of action, and with predictable consequences known.
3
I hope there will be continuing follow ups on this topic. Both the Right and the Left should agree we want the government we have to be as effective as possible and to have a customer focus.
We have to be careful when we generalize about "the government." If everyone is lumped together, incentives to improve are reduced. Find the agencies that are customer focussed and making progress and give them a plug. Find those that are encrusted and making life miserable for customers and expose them. Then, we will start to see progress.
We have to be careful when we generalize about "the government." If everyone is lumped together, incentives to improve are reduced. Find the agencies that are customer focussed and making progress and give them a plug. Find those that are encrusted and making life miserable for customers and expose them. Then, we will start to see progress.
1
I just have to add to thr previous comments regarding 8 text messages, costing $7.
It seems to me a cost of 7 cents should have been closer.
It seems to me a cost of 7 cents should have been closer.
2
I can't imagine that anyone needing to be "reminded" to enroll in college should be in college.
3
Perhaps the problem is with your imagination. Many capable people need to be reminded of deadline dates, especially when time is needed to get information and fill out forms.
President Obama also created the SAVE, Securing Americans Value and Efficiency), program which asked all government employees to submit ideas that would improve government and save money. Many ideas were implemented and saved lots of taxpayer money.
From 2009 to 2012, the best submissions were highlighted with an award for the winner of the best idea for the year. The SAVE program seems to have ended in 2013. Perhaps due to lack of media coverage?
https://www.whitehouse.gov/save-award
From 2009 to 2012, the best submissions were highlighted with an award for the winner of the best idea for the year. The SAVE program seems to have ended in 2013. Perhaps due to lack of media coverage?
https://www.whitehouse.gov/save-award
4
Yes, good point. In fact, SAVE is actually where the double-side printing tweaks and savings ideas originated.
1
The idea is a good one and well-tested over many decades.
Still, I wonder why it costs the government $7 to send 8 texts.
Still, I wonder why it costs the government $7 to send 8 texts.
4
Market power by the mobile company?
Just wait a while and they'll find a way to blame Hillary or these high costs.
If you read the actual report of the so-called Social and Behavioral Sciences Team (SBST), then you will not only see that the nudges barely had tiny impacts but also that the failures in initial program insights were due to bypassing/ignoring business expertise that has long been obvious to those in those fields (whether such expertise is held by private sector or public sector "bureaucrats").
It is interesting that the Times has an editorial at the same time, "Speedy Help for Victims of College Fraud," where the Times defends so called vulnerable consumers (but, as a side bar, are those 40 State AGs offering to put in any funds?), while at the same time allowing the uncritical publication of at least three articles and commentaries on SBST activities, which in some cases represent human experiments on vulnerable consumers, i.e., modifying or even suppressing standard consumer outreach in order to perform a "cool" experiment on how small changes in messaging impact behavior. Meanwhile the SBST members are able to develop academic publications and CVs, while the harmed consumers are just a speck in the rearview mirror of scientific inquiry. Who will go back and help the borrowers who defaulted because they didn't receive at least the standard, basic info? The problem appears to be well-intentioned but meddling political appointees such as SBST, rather than calcified bureaucracy.
It is interesting that the Times has an editorial at the same time, "Speedy Help for Victims of College Fraud," where the Times defends so called vulnerable consumers (but, as a side bar, are those 40 State AGs offering to put in any funds?), while at the same time allowing the uncritical publication of at least three articles and commentaries on SBST activities, which in some cases represent human experiments on vulnerable consumers, i.e., modifying or even suppressing standard consumer outreach in order to perform a "cool" experiment on how small changes in messaging impact behavior. Meanwhile the SBST members are able to develop academic publications and CVs, while the harmed consumers are just a speck in the rearview mirror of scientific inquiry. Who will go back and help the borrowers who defaulted because they didn't receive at least the standard, basic info? The problem appears to be well-intentioned but meddling political appointees such as SBST, rather than calcified bureaucracy.
2
I am disheartened how many people think the government should use so called simple nudges. Subliminal messages may partially work, but I would be appalled if the government used them. Are behavioral nudges necessarily different? Take something as simple as a default choice. The very fact it works is why we should be skeptical of them, at least until we fully understand why they work.
Suppose they work because some people are inattentive. It is not too far of a stretch to think that education and attentiveness are related. If the government mandated a default position for retirement investing, it is a paternalistic tax born by the less educated. I can be persuaded that such paternalism is beneficial. But it seems very dishonest and undemocratic than simply taxing the poor to get them to do what you want. But of course that could never pass. Instead the government can simply use a default and hide behind the ruse that they are not really constraining people. But the very fact that they work suggests that the government is constraining some people.
I do not mean to say behavioral nudges can never be used. But I am worried about adopting them until we understand exactly why they work. There are important philosophical issues that need to be vetted before they are used.
Suppose they work because some people are inattentive. It is not too far of a stretch to think that education and attentiveness are related. If the government mandated a default position for retirement investing, it is a paternalistic tax born by the less educated. I can be persuaded that such paternalism is beneficial. But it seems very dishonest and undemocratic than simply taxing the poor to get them to do what you want. But of course that could never pass. Instead the government can simply use a default and hide behind the ruse that they are not really constraining people. But the very fact that they work suggests that the government is constraining some people.
I do not mean to say behavioral nudges can never be used. But I am worried about adopting them until we understand exactly why they work. There are important philosophical issues that need to be vetted before they are used.
REALITY-CHECK ! ! !=
1) Maybe, they should have checked with an IT person before they did their A/B Test on Printers. The cost of extra sheets of paper is minimal. However, unless you have really rugged, industrial-grade printers, sending duplex-print-jobs to them, creates lots of paper jams, helpdesk calls, and downtime, when nobody can print ANYTHING.
For me, unless I'm using a printer like they have behind the counters @ Kinkos- If I want to print duplex, I print out the odd-numbered pages first, and then flip the paper and print the even page numbers on the back. It's a little tricky, but ultimately faster, and no paper jams.
2) TEXT-MESSAGES?? Are you kidding me?
Does the Government really spend $7 to send 8 pre-programmed text-msgs?
Sanity-Check= Any student who is unlikely to enroll in College unless they get 8 text messages from the government, has ultra-low motivation to attend in the first place, and their probability of being a successful student is ultimately very low.
3) Coerced enrollment in a so-called 'Saving-Plan' is simply that, you pulled a fast-one on them, and in the end, almost all of those people will drop out and pull their money back, after they realize that they have essentially been taken-for-a-ride, a ride not of their own choosing.
4) Tax-Form, I suspect that 'the truth' behind this is more complex than is being revealed.
This is similar to a common scheme from the 70s, put a 6-ft cutout of a cop in a store up front, & shoplifting decreases
1) Maybe, they should have checked with an IT person before they did their A/B Test on Printers. The cost of extra sheets of paper is minimal. However, unless you have really rugged, industrial-grade printers, sending duplex-print-jobs to them, creates lots of paper jams, helpdesk calls, and downtime, when nobody can print ANYTHING.
For me, unless I'm using a printer like they have behind the counters @ Kinkos- If I want to print duplex, I print out the odd-numbered pages first, and then flip the paper and print the even page numbers on the back. It's a little tricky, but ultimately faster, and no paper jams.
2) TEXT-MESSAGES?? Are you kidding me?
Does the Government really spend $7 to send 8 pre-programmed text-msgs?
Sanity-Check= Any student who is unlikely to enroll in College unless they get 8 text messages from the government, has ultra-low motivation to attend in the first place, and their probability of being a successful student is ultimately very low.
3) Coerced enrollment in a so-called 'Saving-Plan' is simply that, you pulled a fast-one on them, and in the end, almost all of those people will drop out and pull their money back, after they realize that they have essentially been taken-for-a-ride, a ride not of their own choosing.
4) Tax-Form, I suspect that 'the truth' behind this is more complex than is being revealed.
This is similar to a common scheme from the 70s, put a 6-ft cutout of a cop in a store up front, & shoplifting decreases
Shhh...don't let the Tea Party Republicans find out about a program that will make the government more effective. They'll want to shut it down.
3
Even worse; do the testing and do the opposite. Anything to push conservative goals upon everyone else
A major challenge with government programs is that once they're established, it is very difficult to discard them even if they don't work.
We have so many overlapping program that work against each other but we can't streamline.
These nudges are great and should be done as much as possible.
We have so many overlapping program that work against each other but we can't streamline.
These nudges are great and should be done as much as possible.
1
This is fine--as far as it goes. But it presents the topic simplistically.
Government operations: bad; academic role: largely irrelevant; business: good as if the field of behavioral economics made no contribution here, or there are no bureaucratic businesses. Has that been your experience when calling a company for technical assistance, when trying to follow the directions in a package?
Efficiency is presented as always good. Issues of ethics, blow-back, unintended consequences, trust are ignored.
The concept of "hard data" is utilized as if there aren't issues of research design, interpretation of what the results are telling you, biases, blind spots, paradigm-limiting perspectives. (Remember the psychology replication articles of a few weeks ago? 60% of prominent studies couldn't be).
Will it work for everything the government does? Should the Pope be doing Google-type experiments when he speaks to us about injustice and climate change?
So a particularly nice addition to the arsenal, and a pleasant if implicit recognition of what government aims to do for us in an era when that has been forgotten. But let's not go overboard. Try to foresee the problems this this one-year-old program might face, and try to head them off.
Do something meta. That is, put this "tweak" itself under the experimental microscope and seek to determine when it works, and when it doesn't.
Government operations: bad; academic role: largely irrelevant; business: good as if the field of behavioral economics made no contribution here, or there are no bureaucratic businesses. Has that been your experience when calling a company for technical assistance, when trying to follow the directions in a package?
Efficiency is presented as always good. Issues of ethics, blow-back, unintended consequences, trust are ignored.
The concept of "hard data" is utilized as if there aren't issues of research design, interpretation of what the results are telling you, biases, blind spots, paradigm-limiting perspectives. (Remember the psychology replication articles of a few weeks ago? 60% of prominent studies couldn't be).
Will it work for everything the government does? Should the Pope be doing Google-type experiments when he speaks to us about injustice and climate change?
So a particularly nice addition to the arsenal, and a pleasant if implicit recognition of what government aims to do for us in an era when that has been forgotten. But let's not go overboard. Try to foresee the problems this this one-year-old program might face, and try to head them off.
Do something meta. That is, put this "tweak" itself under the experimental microscope and seek to determine when it works, and when it doesn't.
1
Nicely put "experiment relentlessly, keep what works, and discard what doesn’t. Following this recipe may yield a government that’s just like Google: clear, user-friendly and unflinchingly effective".
Somehow I feel here at last is a aha moment, an inexpensive technology tool for countries like India being lead by Tech savvy Prime Minster Modi. He is right now in Silicon Valley meeting heads of heads of global technology giants to solve myriad problems of governing extremely diverse country like India, searching for the elusive El Dorado
"Experiment relentlessly, keep what works, and discard what doesn’t. Following this recipe may yield a government that’s just like Google: clear, user-friendly and unflinchingly effective" appears like a magic wand to fulfill cherished dream Mahatma Gandhi who dreamed with passion for relieving every kind of suffering and for wiping out the last tear from the eyes of the last man. .
Somehow I feel here at last is a aha moment, an inexpensive technology tool for countries like India being lead by Tech savvy Prime Minster Modi. He is right now in Silicon Valley meeting heads of heads of global technology giants to solve myriad problems of governing extremely diverse country like India, searching for the elusive El Dorado
"Experiment relentlessly, keep what works, and discard what doesn’t. Following this recipe may yield a government that’s just like Google: clear, user-friendly and unflinchingly effective" appears like a magic wand to fulfill cherished dream Mahatma Gandhi who dreamed with passion for relieving every kind of suffering and for wiping out the last tear from the eyes of the last man. .
1
"Experiment relentlessly, keep what works, and discard what doesn’t. " That sounds a lot like evolution and we all know that evolution is satanic. Just use The Bible as your guide and all will be well. St. Paul tells us to obey our God-Given leaders and do what we are told to do. No messing around with evolving processes when the God-Given systems we have now are perfect.
(Yes, I did major in sarcasm.)
(Yes, I did major in sarcasm.)
9
Testing is smart . But the quality of your tests is ultimately limited by the quality of the ideas you can think to test. Some of these are low hanging fruit, like reminders and simple communication. In the longer run, smart ideas will still be the ultimate source of improvements.
1
The problem is not only the personality of John Boehner, the problem is in the roots of Grand Old Party. Fortunately, it becomes more and more outdated. Of course, we should consider all the points of view, but we should not forget about the reality we live in
"A/B" testing is presented here as though it were some kind of revelatory innovation. It is as old as science itself and its rigorous quantitative implementation is at least as old as the discipline of statistics. Social "scientists" not required.
3
But data analysis IS required. And that is typically very hard to do, especially when you're working out the confidence intervals for narrow variations in percentages. And setting up good randomized experiments takes some skill -- give these people some credit.
I can't tell you how many comments I see that are the usual liberal bickering: some variant of "why are we surprised that 'x' happened..."
I can't tell you how many comments I see that are the usual liberal bickering: some variant of "why are we surprised that 'x' happened..."
2
Funny, looks like the usual conservative/libertarian bickering to me...
1
A couple of problems here:
One, I think these things are moving targets. A new popup telling you to print two-sided might work the first few times you see it, but when it becomes part of the background noise you might start to ignore it.
Two, small percentage point gains might be great at the scale of Google or the federal government, but on the scale that we actually live on - say writing emails to coworkers, those differences probably don't matter.
In addition, if A/B testing consistently yields _functional_ results, then the aesthetic of big organizations will become about efficient function. It seems likely that those those giant functional organizations that care about small percentage point changes will start to all have very similar designs, aesthetics, and operations developed with A/B testing nudging it all in the same direction. Those organizations will soon risk feeling like they don't have a unique or human voice. Which is maybe no shocking thing to say about the federal government, but consider the recent change to Google's logo.
One, I think these things are moving targets. A new popup telling you to print two-sided might work the first few times you see it, but when it becomes part of the background noise you might start to ignore it.
Two, small percentage point gains might be great at the scale of Google or the federal government, but on the scale that we actually live on - say writing emails to coworkers, those differences probably don't matter.
In addition, if A/B testing consistently yields _functional_ results, then the aesthetic of big organizations will become about efficient function. It seems likely that those those giant functional organizations that care about small percentage point changes will start to all have very similar designs, aesthetics, and operations developed with A/B testing nudging it all in the same direction. Those organizations will soon risk feeling like they don't have a unique or human voice. Which is maybe no shocking thing to say about the federal government, but consider the recent change to Google's logo.
2
The government doesn't need to have a unique human voice and it's silly to be anxious about it losing its identity. Making more efficient government is good for everybody. I honestly don't know what point you're making.
3
I made a similar distinction between nudging as a bunch of tips and tricks from the literature, and understanding what you're doing in this blog post.
http://clubtroppo.com.au/2014/12/16/tips-and-tricks-or-tips-and-tricks-o...
http://clubtroppo.com.au/2014/12/16/tips-and-tricks-or-tips-and-tricks-o...
1
It's good to learn of these improvements. I don't know of any examples of business doing this kind of experimentation to help their customers. Businesses only look for ways to charge more for less. Which is why leaders who see government as a means to help citizens are better than leaders who only help business.
1
Hooray! It's about time! I've often wondered over the years why our government agencies didn't partner with some of the best and the brightest in American industry to improve services and save money. Companies do this all the time - why can't government? Why didn't the folks in charge of Obamacare hire Google to help them rollout the website? It would have worked spectacularly from the get-go instead of being a nightmare. As a taxpayer, I WANT efficient government and I'm sure Apple, Google, Microsoft, and other leading U.S. technology companies would be delighted to help make government more effective. The best way to stop people from trying to dismantle government is to make it work!
13
I would say that the government should employ a lot of "housewives," if they want efficiency and organization. Space saving, time saving, money saving are all the bailiwicks of the successful domestic engineer.
2