This ran yesterday in TechPresident. Since then I’ve gotten a number of favorable responses, including several that suggested the government should actually consider having Intuit design the site.
As I thought more about it, it seems to me that this approach — of a process in which muncipalities and states wouldn’t have to make any calculations, but simply input the data, at which point it would be automatically tagged with metadata and the calculations would be done in the cloud, might actually be a good model for all federal reporting processes because of the combined value of:
* being able to capture the numbers as structured data that could be used in so many ways (it’s vision of data-centric organizations that’s in book)
* probable reductions in reporting errors if you had a cloud-based interactive process including numerous checks
* and — actually just thinking of this for first time as I write — creating a Dutch Taxonomy Project-style unified reporting process that could radically reduce the amount of time and money local & state govs. would have to spend to report to feds, just as it does with Dutch companies saving on reports to multiple gov. agencies., while also facilitating collaborative analysis by multiple federal agencies (and transparency!)
What do you think?
What Recovery.gov could learn from TurboTax
By Neal Hannon & W. David Stephenson
With April 15th looming, we’ve been renewing acquaintances with an old friend, TurboTax.
That got us thinking.
New US CIO Vivek Kundra likes to ask why he can do certain things in his personal life that can’t be done by government – and then he makes those same services available from government. We think TurboTax might give him another of those parallels: why not make reporting of expenditures under the stimulus plan as simple for communities as it is for us to prepare and file income taxes using TurboTax? The advantage of a pre-formatted reporting program with a strong set of business rules running behind the scenes (which is, technically speaking, what TurboTax is) is, as the ad says, priceless in a situation where errors can be costly.
After all, there’s a lot riding on accurately recording expenditures on the Recovery.gov site. The Obama Administration wants to prove that stimulus money is being well spent, and particularly that it is meeting the primary goal of saving or creating jobs. Watchdog groups and the media are also eager to see if there’s been waste.
A lot of time and planning has gone into technology for the Recovery.gov website, where all of the reports on stimulus spending will be aggregated. However, the biggest potential for errors is at the beginning of the process, when recipient communities have to report how they spent the money and the results. If those numbers are off, every succeeding step in the process will be skewed.
Since many of those reporting on the expenditures will be small communities without a lot of staff (and even big communities may have lost a lot of their specialists who could do this function), the reporting process needs to be designed so people with no special training or resources can still complete it quickly and accurately. Doesn’t that sound like the challenge non-accountants face every year with doing their 1040’s?
TurboTax, in our estimation, does a great job of mediating between the individual with no tax skills and the complexity of the tax process. Many of its features were designed to address parts of the tax filing process that directly parallel those facing reporting on stimulus spending.
For example, while it is still possible to buy a Turbo-Tax CD-ROM, most people now access it and complete the process online. That’s also critical for stimulus reporting: why set up an expensive and extensive paper-based or desktop-based process when a single cloud-based process both simplifies the process and allows officials to devote more time and money to designing a single, comprehensive, best-in-class reporting process?
When you first log on to TurboTax, it asks you to provide some basic demographic information. That in turn determines whether you fall into several classes where specific deductions might apply or additional information is needed, such as the elderly, those with education expenses, or the self-employed. Since Recovery.gov must apply to every community from tiny ones to New York City, quickly assessing what major categories a specific community may fall in (such as urban/rural, high unemployment/minorities/immigrants) so that it can quickly triage communities in terms of the level of detail and complexity in reporting is critical.
However, we believe the real benefit of a TurboTax-style approach to Recovery.gov would come in the heart of the reporting process. TurboTax asks for data, and automatically generates forms based on input: the user doesn’t have to make any calculations, reducing the risk of error. Likewise, an entity reporting stimulus spending in a rules-based format need not worry about anything but accurate data inputs. Reporting can be automated at the recovery.gov website.
Behind-the-scenes, the data would be augmented, or “tagged” with meta-data (“information about information’), using a structured data format such as XBRL, the eXtensible Business Reporting Language . Once a specific piece of data has been entered into the reporting system and is tagged, it becomes what you might call a “data nugget,” with context and meaning garnered from the metadata. The data lives not as a form, but as information accessible by any computer program that can “call” or retrieve the data. On the Federal level, the FDIC already uses this approach to capture quarterly information from the nation’s over 8,000 financial institutions.
A data-centric approach from start to finish in Recovery.gov is critically important. Collecting accurate data, as opposed to paper-based reports, provides increased accuracy and flexibility in both the collection of data and the reporting of the data. . If structured data is used, it could be accessed both instantly and simultaneously by a number of government agencies and externally by the public to assure accountability and transparency.
But we come back to the issue of accuracy in reporting that individual item of data. That’s where the TurboTax example is again illustrative of what Recovery.gov should do.
First, there’s the chance of a simple entry error, so the respondent should be queried and asked to enter the number again as an accuracy check.
More important, what if the question was misunderstood, or had subtleties that might not be immediately apparent? Think about when you fill out an item on TurboTax that is frequently misunderstood. TurboTax gives you options including watching a brief video, reading the exact wording of the tax code, using its wizard process that guides you through the process, and (a recent addition) also lets you pose a question to online users (via a wiki) who might be able to clarify the answer for you.
Variations on all of those options can and should be available to those filling out Recovery.gov forms online. When Kundra was the chief technologist for the District of Columbia he made extensive use of brief YouTube videos to help employees understand new processes. It would be easy to create an online wiki for other stimulus recipients to ask and answer questions. It would be easy to design a process that would guide the recipient through the various steps of deciding what to enter based on an interactive Q & A process.
Finally, of course, TurboTax runs you through a final review and error catching process, based in part on whether your answers are exceptions to widespread norms (that doesn’t mean that the answers are necessarily wrong, just that they deserve more scrutiny on your part because IRS reviewers are also likely to scrutinize them more carefully). Why not a Recovery.gov variation on the theme?
TurboTax has radically streamlined the 1040 preparation process for many Americans. A parallel, cloud-based reporting system capitalizing on the many years of experience with the online tax preparation process could help Recovery.gov simplify the stimulus reporting process, capture structured data “nuggets” that could also be used in a variety of other reports, and significantly improve the accuracy of the stimulus reporting process.
Now back to your 1040s…
Neal Hannon is senior consultant, XBRL Strategies, Gilbane Group, Cambridge, MA. W. David Stephenson is principal, Stephenson Strategies, Medfield, MA.
Very good! William Eggers had a chapter in his 2005 Gov 2.0 book called “G2B: The E-Gov Invisible Hand” that talks about using process systems to intermediate regulatory compliance. It’s a revolution that’s past due.