Everybody Get Your DUNS! And Why Grants.Gov Needs An Extreme Makeover.

I have no doubt it seemed like a good idea at the time.

The official OMB Guidance to Federal agencies on how to handle stimulus money requires everyone to go through the federal grants portal Grants.gov. Given that the same guidance also requires agencies to coordinate with one another to further the broader interests of the legislation, to streamline things for applicants and grantees, and to track money disbursed under the American Recovery and Reinvestment Act (ARRA) from the moment it leaves the Treasury to the moment it is spent on a shovel, this would appear to make perfect sense.

So, being an independent contractor these days, I decided to try to see how easily the system worked. Surfing over to grants.gov, I see nothing directing me to ARRA, stimulus, or whatever. But that’s OK. Since I know I’m going to need to get registered, I click on the get registered link where — still no specifics about ARRA — I must now choose between registering myself as an organization or as an individual. OK, lets go with “organization.” Here I hit my first roadblock:

Step 1: Obtain DUNS Number

The DUNS number is issued by Dunn & Bradstreet, and appears to be something of a universal identifier for government purposes. Why the government outsourced this function is probably lost in the mists of time, but OK, whatever. Happily, Grants.gov has a link to the Dunn & Bradstreet site to apply for a DUNS Number. This includes the helpful information that, for some reason, Dunn & Bradstreet is a bit backed up at the moment as lots of people are applying for DUNS Numbers. As part of filling, I discover I first needed to figure out my Standard Industrial Classification (SIC) Code. Again, a helpful link takes me to the right website, so all I need to do is figure out if my new business is “consulting” or “business consulting.” so it’s just fill out the rest of the form, submit, and wait.

Now that I have my DUNS Number (did I mention there may be some wait, as they are backed up at the moment), I can proceed to the next step: registering in the Central Contractor Registration Database (CCR Database — and no, they do not provide the lyrics to Bad Moon Rising or any other song). After that, I’m almost ready start applying for actual stimulus money . . . .

I learned two things from this. First, anyone who thinks they might, possibly, perhaps, vaguely could someday want to apply for any ARRA money should go out and get themselves DUNS number so they can get registered in the CCR ASAP. And, to its credit, the OMB guidance said only about a zillion times that agencies needed to go out and evangelize to prospective grantees (especially little ones) to go get DUNS Numbers.

Second, and more importantly, Grants.gov needs an extreme makover — quickly. The idea that I need to get a number so I can register for another number to go into a database that will be redundant with other databases is rather ridiculous in this day and age. worse, it creates a serious barrier to every single desired outcome. The current systems, as it stands, not only makes it hard to apply (especially for small organizations or folks who find out about the procedures at the last minute), it makes it needlessly difficult for the Feds to track the ARRA money.

My recommendation below . . . .

As the Washington Post recently observed, the Obama Team keep running into the unexpected brick wall of finding that the Bushies did little to upgrade the government technology they inherited from the Clintons. So they keep assuming they are going to be able to do all this stuff which is utterly routine these days, only to discover that before they can launch they must simultaneously invest in needed technology and decide whether or not to build on the primitive, dead-end stuff they inherited.

And so it goes for Grants.gov. I have no doubt this was cutting edge back when they put it together in 1999. And I expect that it’s mission is further complicated by the need to comply with a whole host of laws around giving out government grants and maintaining federal databases, each developed by its own separate agency and using its own quaint rules and definitions.

Which is why I recommend that Team Obama ditch the existing package and start from scratch. Don’t try to build on Grants.gov. Create an entirely separate system for ARRA designed to meet the statutory goals of making it easy to apply, easy to use, easy to track money, and easy to define and collect metrics that show whether this stuff actually accomplishes anything. Oh, and for bonus points, make it customizable in a way that will let people tag information so they can use it for data collection and research purposes we haven’t even thought of yet.

Happily, all of this stuff is off the shelf technology these days.

Really, all OMB (or Joe Biden’s oversight committee, or whoever is centrally coordinating this mess) needs to do is build a fairly simple database with standardized fields for the desired inputs. This is what the standard forms at Grants.gov do anyway, but on paper and requiring significant and painful reentry. Why not, for purposes of ARRA, start from scratch? Don’t use DUNS or other private sector tracking info. Develop a separate “ARRA number” generated by OMB itself (not that hard). Ideally,this one number would be sufficient, and we wouldn’t make people who have no interest in being “federal contractors” other than for purposes of ARRA get into the general database.

Then develop our own ARRA database. Start with a basic form for the metrics OMB wants to track everywhere. Then set aside a section for each agency administering an ARRA program. Then — through the magic of “cloud computing” — simply have each applicants submit the information needed online. OMB could track all the information immediately, as could any individual organization to the extent permitted by the Feds. If we include an ability of individuals to voluntarily tag the information, it will enhance the searchability of the “ARRA database” for a variety of metric tracking purposes.

This also solves the “how will poor little organizations keep track of all the money” problem that has folks clucking like chickens that only big, sophisticated players can handle the book keeping accounts. Anyone who can handle the bookkeeping needed to run a business can do simple bookeeping for ARRA online on a secure database maintained by OMB. Even better, this allows OMB to use the same tracking number for ARRA money from the beginning of the process to the end of the process, AND it serves to act as a check on comingling of funds and other games used by sophisticated players to hide various forms of peculation.

Yes, there will be tradeoffs for things like spam and security. I have no doubt folks can come up with a list of possible horribles and problems with my two second design here. But please observe that these problems have already been solved a gazillion times over. Furthermore, based on the existing statutory requirements, OMB and the rest of the folks at the center already need to solve the security and privacy problems. All I’m talking about is taking a step back to think of this in terms of rational process.

Similarly, I expect real database designers and folks with real experience working with cloud computing, social networking, and other related areas of expertise will come up with numerous improvements over my proposal. At least I hope so. all I'[m trying to do is focus on the fact that if we take a step back and think about what we actually need, rather than rush to fit everything into pre-existing molds, it may turn out to be easier, cheaper, and all around better to build something from scratch to deal with ARRA and plug it into the Grants.gov portal rather than break this job up into a million needlessly complicated pieces working at cross-purposes with each other.

Which brings me to the last potential objection — time. OMB has set aggressive deadlines for agencies to have applications ready at Grants.gov. But given the nature of what is up at Grants.gov now, it may actually take less time to build an ARRA-specific back-end and plug it into the website. Or, to quote Friar Lawrence from Romeo and Juliet: “Wisely and slow, they stumble that run fast.” Alternatively, even if they can’t get my proposed extreme makeover ready for initial applications, they can develop it in time for use to collect information on metrics and to track ARRA money after granting applications.

Nothing I’ve proposed here requires developing anything new. It only requires money, expertise, and a willingness to exploit the potential of the technology. For once, Congress has appropriated a nice piece of change for the project. The Obama people have certainly shown an appreciation for technology and an ability to make it work for them in sophisticated ways. Hopefully, they will take a minute to put these elements together. If there is one thing we’ve learned in the tech trade, it’s that sometimes it’s easier to start again from scratch than to keep trying to make marginal changes and upgrades. Now would be a good time to apply that lesson.

stay tuned . . . .


  1. Welcome to the Public Sector! Isn’t seeing the execution of Public Policy you used to advocate fun?

    As to starting from scratch. Well yeah, but that requires a contract to be let, well that requires a DUNS……

  2. JohnMc:

    It’s a lot more fun than seeing it done by the private sector, that’s for damn sure! This looks like a bad user interface, hastily assembled, to solve a complex problem. What we had before was a Mafia bust-out.

    Bad user interfaces can be fixed. Following Harold’s suggestion, without too much work.

    Embedded mafias, such as the Bush administration and the whole financial sector, are a much more difficult matter. Which is why having Tim Geitner in charge of the Treasury is so very un-reassuring to me.

  3. JohnMc:

    Actually, I enjoy troubleshootng the process all the way through. The fact is that government, like industry or any other big generic generalization, can be made to work.

    Nor do I think what I’m suggesting requires a contract. I understand the government employs people with the necessary skills, or could hire them reasonably quickly. What is important is for those implementing this to occassionally take a step back and ask “do I really need to accept things as they are, or can I rethink this and do it better.”

  4. I get flashbacks.

    My wife just went through the process, but it was 10 times worse. Because she had an old, partially completed record…

    And, for most of her government contract work, she uses her old married name. Because you can’t easily change names… The dude’s been gone 5 years.

    You should give a call some time, and talk to her. But not today. She’s in DC today, at a conference, designed to “make the application process easier”. đŸ™‚

  5. Mark S:

    Please have her contact me, I would be very interested in talking to her.

  6. Um, I thought the new leftist agenda wanted to avoid no-bid contracts? And that things executed by civil service tend to be kinda expensive, since you have to pay for all those non-fireable non-working workers? And then pay for civilian consultants who actually write the program and maintain it.

    Software is not just a matter of finding the programmer to write it, it’s a matter of user interface design, which takes more face time with the people who already have the process to be automated. Designing a new process is just as time-consuming, if not more so. Then you have to document the system, and test it out before deployment, and maintain it over time, etc.

    There’s a reason software companies are so large. We have, I think, 10 people at IBI per programmer, to test, document, maintain, help users, be our own MIS department, accounting, finance, customer education, customer consulting (to help set up systems), sales, etc. Some of that would not be necessary in a government operation, such as sales and accounting, but support and maintenance do need to exist for any system.

    So we have a procurement process, and it takes years to assemble bids, and get them approved, and then to start writing to a spec written at the beginning of the process, even though technology may have moved on in the interim. On the other hand, keep it all in-house, and you have to hire a lot of people to do anything, and there will probably be more opportunities for shenanigans. Governments put these bidding processes in place for a reason. Think Tweed Courthouse, or Blackwater.

    Or, why are [NY] City Hall’s computer systems always about five years behind the times?

  7. Alternatively, your post reads like you said “Programmers are bright people, they should be able to toss something together in no time, and have it work right out of the box.”

    I could just as easily say “Policy wonks are bright people, they should be able to toss a national telecom policy together and have it work right out of the box.”

    Just as what you do is hard, what we do is hard.

  8. Jon:

    You misunderstand my point. I am simply saying that my proposal requires no new technology or development of something radical. I confess, I sometimes think people reading these posts look for ways to be offended. Or perhaps you simply have no concept of how many people would read what I describe as something impossible and overly complex rather than something readily designed and implemented — albeit with the same level of hard work and skill required for any other significant project.

    There is no need to do this with a contract. The federal government should actually hire qualified people so it is capable of carrying out its functions, rather than looking to outsource everything — especially in the IT area which government should increasingly utilize. The effort to construct what I am talking about, preferably using available tools and off the shelf technology as much as possible, is non-trivial but manageable.

    Which, I suppose, IS how I feel about media and telecom policy. Many of the mistakes I point out above (doing incremental tweaks, accepting the world as is) are common in the field and I spend a good deal of my time trying to get smart people with solid skills to step back and consider how to do the job effectively rather than just go haring off after some convenient detail.

  9. Fine, don’t do it with a contract. But something like this needs to be robust, which means a lot of testing time and server time, it needs coordination between the different agencies that dispense and collect the information, which gets into penetrating interpersonal and inter-agency rivalries, it needs security, it needs to be designed and put together consciously. Those agencies may themselves be using incompatible systems, for which interfaces need to be written. Someone would need to be responsible for handing out ARRA numbers.

    Why not just use Tax IDs? Every person or corporation with a bank account has to have one. Oh, right, that way lies more possibility of financial shenanigans. You see? Every simplifying assumption needs to be examined.

    Without a procurement process it wouldn’t take five years, but it would still be non-trivial to replace from scratch. You’re talking about replacing what we have NOW with something BETTER RIGHT AWAY, and I’m telling you you won’t get it, even under good circumstances, for months if not years.

    Criticizing X for being bad is easy, constructing X’ that is actually good, takes talent and a lot of work.

    That the technology for writing nice forms and managing databases exists, makes it the work of months rather than years, but rolling it out Right Now is well-nigh impossible.

  10. So don’t do it? That is as much the logical extension of your statement as you seem to be reading mine.

    To make clear, again, of course putting together a big project in a robust fashion takes serious resources, planning and time. Although having dug around a bit before posting this my impression is there is more out there to cobble this together than you think. And, as I said above, it may need to be phased in, with some elements constructed over time. Certainly I wish that the ARRA deadlines were more sensible and less politically driven. but none of this is an excuse for trying to simply layer stuff on the existing system. Because, by your logic, you can NEVER replace outmoded or bad technology. It will always take too long or conflict with other priorities.

  11. I’m having flashbacks to a book I worked on last summer:


  12. Harold,

    Pretty interfaces aside, its like paving pot holes on a bridge that has too many bolts missing. The core problem is that the grants process is still a series of independently developed schemes by the individual government depts (FCC, Comm, Ag, etc.)

    The various process mavens in each need to agree on a ‘grand unifying theory’ that they all can accept with supplemental components based on grant specifics.

    I have seen kludge in SW design and it can get ugly.

Comments are closed