Tag Archives: Vivek Kundra

The downsizing of data.gov: learning to manage expectations

Tim Berners-Lee at TED giving the famous "Raw data now" talk

Bad news for open government and open data activists everywhere. data.gov, the Obama administration flagship open data portal, is being taken down downsized on the wake of the recent federal budget cuts. So is former White House CTO Vivek Kundra’s IT dashboard. Especially data.gov is a hard blow: that is the template, imitated by the faster-moving governments and envied by the citizens of the slower-moving ones.

What went wrong? Steve O’Keeffe summarized it like this: GIGO and overpromise. I am not qualified to judge GIGO in this case, but overpromise rings a bell. The movement is wired along Tim Berner-Lee’s “Raw Data Now”: release the datasets, civic hackers and market forces will do the rest. Decision makers in the faster-moving administrations were all too happy to comply: the number of datasets you release is an easy to monitor measure of acitivity, and it looks great on press releases.

The demand-side has been sagging. This is unsurprising: interpreting data to tell causally convincing stories is hard. Civic hackers need to know their measurement theory, probability theory, statistics, econometrics; computer prowess does not take you beyond download. We did have some good examples of data driven journalism, but most of the media ignore the data and stick to interviewing academia and gov brass when they want coverage of economic/social/environmental issues. Makes sense too: there is not enough readership for data driven journalism yet.

Worse, we were told that new ecosystem of innovative services would arise from the availability of government data, leading to growth and jobs. Hard to resist: the package of innovation, growth and jobs in one phrase is one of the very few passwords that will unlock serious funding these days, and proponents and funders alike went with it. Couple of years down the road, we do have some cool apps and some companies that use the data. We even have some jobs generated around data availability, but the numbers are unimpressive. The most I’ve heard trumpeted is 60 employees for a single company. That, too, is hardly a surprise: if your business is based on an open-access, unexhaustible resource like gov data, economic theory tells us it’s going to be hard to bake any seriouse margin into it. You tend to get undercut and outcompeted by non profits and zero-overhead college students operating out of laptops. Profit requires scarcity, not abundance – just ask music recording studios. Keefe’s post contains an interesting little fact, and that is the private business has indeed invested in data, but private and very muck locked down.

Given all this, and in the light of the demise downsizing of data.gov, I would recommend the open gov movement to resist the temptation to promise anything we are not sure we can achieve in any scenario. Envision low-cost, low-hype operations; offer the collaboration of nonprofits and the civil society; emphasize that, while people are welcome to make money out of value-added-on-open-data services, that is not the point of the exercise. The point is increasing the transparency, accountability, and efficiency of public policy. It will be less cool, it will take us off the spotlight and the big funding grants, but it will keep the movement going, almost invulnerable to disenchantment, budget cuts and lobby capture. If I am wrong, great: another year from now, we can make a comeback and boast all the jobs open data will have created in the mean time.

Obama’s data and Clarke’s first law

As was expected, Wired’s June issue has a story about Vivek Kundra, the first-ever White House Chief Information Officer. The Obama administration’s vision on federal data release in a machine-readable, user-rateable and taggable form is indeed very fascinating.

Towards the end of the interview, almost as a side to the main topic, the interviewer asks this question:

As CTO of Washington, you moved tens of thousands of employees from Microsoft Office to Google Apps to save money. Part of your new agenda is shifting the government to cloud computing and using free software. How will that happen?

I had missed it: in October 2008, as Obama was busy with his campaign, Kundra worked as Chief Technology Officer of District of Columbia. In that capacity, he moved 38,000 employees of the District’s administration from Microsoft Office to the Google cloud suite, obviously saving quite a lot of taxpayer money.

The problem with this story is that it looks like a theoretical impossibility. Incrementalist thinking – Yale’s Charles Lindblom being its most prestigious academic – is extremely influential in the American political science tradition. Incrementalists looked up deep, important reforms like the introduction of the Federal budget under the Roosevelt administration, and concluded that reform happens in small steps, aiming for what is possible, i.e. “the present situation plus or minus five per cent”. Migrate 38,000 employees from a system to another, mister Kundra? What about trade unions? And suppliers? How can we train employees for the new system? How can we get the middle management’s consensus? Trust us, mister Kundra, it can’t be done. Let’s do a pilot project instead, or a feasibility study. Something more… realistic. Incremental.

The incrementalist position – which I respect deeply – remind me of Clarke’s first law: when a distinguished but elderly scientist states that something is impossible, he is very probably wrong. Did Kundra ever read Lindblom? Did he know his reform was impossible whan he went ahead and did it? He made it happen anyway. I wonder whether a little less of scientific realism and a little more healthy recklessness would not be a better recipe for reformers all over the world.

I dati di Obama e la prima legge di Clarke

Com’era prevedibile, Wired di giugno intervista Vivek Kundra, il primo Chief Information Officer nella storia della Casa Bianca. La visione dell’amministrazione Obama sulla distribuzione dei dati in forma machine-readable, taggabile e valutabile dagli utenti è effettivamente un argomento superinteressante, che tutti quanto faremo bene a tenere d’occhio.

Come CTO a Washington hai spostato decine di migliaia di dipendenti da Microsoft Office a Google Apps per risparmiare. Una parte della tua nuova agenda è di spostare il governo verso soluzioni di cloud computing e di software libero. Come farete?

Mi era sfuggito: a ottobre 2008, mentre Obama era occupato con la sua campagna, Kundra lavorava come capo della tecnologia del District of Columbia. Da quel ruolo, ha spostato i 38mila impiegati dell’amministrazione distrettuale da Microsoft Office alle applicazioni cloud di Google, ovviamente risparmiando un bel po’ di denaro del contribuente.

Il problema di questa storia è che, dal punto di vista teorico, la mossa di Kundra non è possibile. Nel pensiero americano sulla Political Science è molto influente il pensiero incrementalista, di cui Charles Lindblom è probabilmente l’esponente più noto e prestigioso. Gli incrementalisti hanno studiato riforme profonde e importanti, come l’introduzione del bilancio federale di previsione da parte dell’amministrazione Roosevelt, e hanno concluso che le riforme si fanno a piccoli passi, mirando a ciò che è possibile: “la situazione iniziale, più o meno un cinque per cento”. Migrare 38mila impiegati da un sistema all’altro, mister Kundra? E i sindacati? E i fornitori? Come risolvere il problema della formazione? Come ottenere il consenso del middle management? Dia retta, non si può. Facciamo un progetto pilota, uno studio di fattibilità, una cosa più… come dire? Realistica. Moderata. Incrementale.

Le argomentazioni degli incrementalisti – che rispetto molto – mi ricordano la prima legge di Clarke: se un anziano e famoso scienziato dice che una certa cosa è impossibile, probabilmente ha torto. Kundra ha mai letto Lindblom? Sapeva che la sua riforma era impossibile? Comunque l’ha fatta. Mi chiedo se un po’ meno di scientifico realismo e un po’ più di sana incoscienza possano aiutare anche i riformatori, in Italia come in tutto il mondo.