Measures and measuring systems as a transparency tool #big #data #policy #behavioural
I wrote yesterday about my number-crunching experience:
Dating and a cup of coffee #BerlinDiaries FB https://www.facebook.com/notes/roberto-lofaro/dating-and-a-cup-of-coffee-berlindiaries/10151624917997786 Frype http://www.frype.com/robertolofaro/blog/?p=8872662
If you want more details, you can have a look at robertolofaro.wordpress.com (a work in progress- but it will take time!).
With this article, whenever it makes sense I will post on both my blog (replicated for continuity) and my main online profiles (Facebook, but you need to login, and Frype, if your are not on FB; you can see the “likes”, but the number of readers is visible only to the author).
As you maybe know, Italy is going to have an election cycle in late February- a couple of months ahead of schedule.
Less than two months- but it could be two aeons (eons, for my American friends).
Sometimes, I toyed with the idea of a telenovela set within the Italian Parliament.
I know that when a pornstar was a member of the Parliament, a movie was made- but that is not my model.
There are so many twists and changes and drama in just a single week within Italian politics and its surrounding areas, that any telenovela can be considered boringly routine, “calma piatta” (dead calm, was the original title of the movie, I think).
Well, Italian politics seems picked up from classical Greek theatre, and the only constant is that “never” means “maybe”, verging on “probably”.
Routinely, Italian politics has a crisis- and despite being a mainly Roman Catholic country, we have the curious habit of exacting the modern equivalent of human sacrifice, exposing the entrails, and reading the future.
Routinely, Italians are fond of statistics, numbers, and “German” administration trends- introducing yet another commission or organization focused on cutting costs, restructuring bureaucracy, and overall becoming “un paese normale” (an ordinary country).
I heard this speeches since the first election campaign that I observed through my family (I was 7 years old), and nowadays are not even stale: a political ritual, a routine that has to be done.
The funniest moment? When Mr. Berlusconi was a sitting Prime Minister (formally: President of the Council of Ministers), and his challenger, representing the Centre-Left but formerly with the then-defunct Christian Democratic Party, was Mr. Prodi.
In a public TV debate, suddenly it sounded as if the roles were reversed.
And also the current campaign is shaping up in a funny way: as it seems that everybody is stating that they will change the most recent structural innovations introduced by the current technocratic government… including the political parties that supported it, and even the sitting Prime Minister, Mr. Monti.
So, never say never in Italian politics: everything is temporary.
Except “temporary taxes”- those become permanent; I remember that when I was a “cadre” in the early 1990s, I had a deduction on my salary, to finance for building houses for cadres.
With a glitch: they had stopped building long before I started paying the tax… and the same applied, as far as I know, for a significant number of “temporary taxes”, e.g. those embedded within the hefty price paid at the pump by Italians.
Moreover: each time there is a change, transparency is promoted and measured- but quite often the measurement is applied to the bottom line, and there is enough “fat” within the Italian administration that most can adapt, and comply without really changing.
Yes, the “bisogna cambiare tutto affinché nulla cambi”- everything has to change, so that nothing changes.
But I must confess that I saw that also in my activities within companies, after in late 1980s, and moreover since the 1990s, companies became obsessed with “measuring”, “KPIs”, “Six-sigma”, and other quantitative initiatives.
Incidentally: it is a chicken-and-egg, as the quantitative approach became appealing when… personal computers started to have Multiplan, Lotus 1-2-3, and then the jack-of-all-number-crunching-trades, Excel.
I am currently digging in detail into Excel 2010 while preparing for the 2013 version: and finally it seems able to replace most of the number-crunching and number-presenting softwares that I helped to sell in the past.
So, what do you do when technology enables you to collect more data? Invent ways to do analysis using more data.
Nowadays, you probably read about “Big Data”- in the past, it was “data mining”, as it was done exceptionally, due to the cost of resources, while now the aim is to do it continuously.
Back down to Earth and away from technocratic lingo: it means that in the past once in a while your supermarket chain was rummaging through your sales receipts, matching with the payment methods and, if available, the loyalty card.
Then, the results were used to both target you with specific offers, and to maybe change the layout of your supermarket, promotions, orders to their suppliers, and “assortment planning”- i.e. which sizes, shapes, brands to carry (and maybe which ones could make sense to replace with their own brand).
In the near future, it might well be that when you use that handy barcode scanner to do your grocery shopping, your supermarket will instantaneously text message you with a discount coupon for a special offer that complements what you already scanned.
Or text message a recipe that requires few more ingredients.
Not too long ago, in London some shops started to text message customers when they were nearby, but that was still done as an afterthought, while now it could be done dynamically.
How does this influence politics? Well, if you read articles about the recent Presidential Campaign in the USA, a lot.
As having an existing customer base, collected in the previous campaign and nurtured through connections via social networks was useful to seed it- e.g. by using existing contacts as a beachhead in each community.
Italy since at least Spring 2011 is undergoing a significant change in the concept of “transparency”, as even before Mr. Monti took office a new law basically removed banking secrecy.
Enter “Big Data”: as I discussed in December 2011 in a previous post on my blog, the concept is that each transaction on each account could potentially be automatically notified to the Italian State.
If you consider that in Italy cash registers are, by law, required to issue a coded tax receipt, and that Euro banknotes are uniquely identifiable (and that Italy lowered the amount that can be paid cash), it requires only a small technical change to have each cash register equipped with a banknote checker that reads the banknote identification- and notified that as well.
Ok- maybe when the broadband Internet will be set up, but the potential is already there.
So, what could be a “Big Data Government”?
Beside Orwellian fears, it could actually adjust its intervention based on the actual financial flows: it is sheer lunacy to exact upfront payments on future income from businesses that have no cash coming through, as this implies pushing them into debt, further reducing the resources available for development.
And this brings back the title: “Measures and measuring systems as a transparency tool”.
In the late 1990s, a terabyte database was unusual, and even retail companies had to consider how long they would keep data.
So, at a certain level of detail, you could have day-by-day information, but further down, e.g. within the assortment, you had to cut down to no more than weekly or even monthly aggregate data (summaries).
Therefore, your own receipts could maybe have to be kept for legal reasons on paper, but storage was expensive, and their digital copies maybe disappeared within a week- or even within a day (I saw other cases as well- the larger you are, the more resources you have to invest on keeping being larger than your competitors).
Moving away from taxes and supermarkets, let’s see what happened when you introduced “Key Performance Indicators”.
In many cases, those were revised (if at all) quarterly, or even each year, while being produced monthly or even daily.
Why? Because each revision implies training, maybe changing activities, and so on: an expensive exercise.
But the longer you kept a KPI (or any other measure) in place, the more you gave incentive to adapt to the measure, not to the intended result.
As an example: if you say that as a relationship manager I cannot invite prospective customers to lunch if the lunch costs more than 20 EUR, it could make sense, if the customer is worth 100 EUR.
But if the customer is worth 10,000,000 EUR of already acquired business? The alternatives are a) I pay for a meal out of my pocket b) I do not invite the customer to lunch.
The third and fourth alternatives that I willingly ignored? Cheating (i.e. spreading the cost across multiple fake receipts or recovering in some other ways), and… inviting your multi-million EUR customer to McDonalds.
Usually, the result in those cases was that eventually the relationship manager migrated elsewhere- bringing the business to another organization.
With “big data”, you could actually increase and dynamically adjust the measures, but would that improve the situation?
No. It would just require better skills at adapting.
Without investing in mega-databases, a simple preliminary approach is to prepare measures (KPIs, reports- pick your choice of words, methods, technologies) that actually require to change the behaviour.
A favorite story is from my early 1990s experience in retail, while I was shipped to do a “icing on the cake” consulting mission from my CEO to a large retail chain, who had got our HR and payroll system.
My “icing on the cake”? See how “optical discs” could be used to reduce the paperwork.
At the time, optical storage wasn’t legally allowed, and therefore the aim of my study, with the support from a legal specialist on payroll, was to show how it would benefit both the legal authorities and my customer to use a centralize optical disc that contained all the information.
While preparing the study, as it was a point about the discussion (saving money while keeping data for the legally required timespan), we were told a story that exemplifies how a measure can influence behaviour.
At the time, the legal authority storing and processing all the information about payroll and retiring benefits was said to have the largest data centre in Europe.
Nonetheless, it was before data travelled electronically: all the information was to be sent on magnetic tapes.
You can imagine: a gazillion of magnetic tapes, that had to be kept by companies for 10 years, and by the organization… for how long it could be considered safe.
Usually: until the data represented people who were not receiving anymore retiring benefits- which, with the Italian system, could imply until any survivor stopped receiving the retiring benefits of a spouse, etc.
With a glitch: Italy has always had a quarrelous Parliament, hell bent on changing laws to cover for specific constituencies- at a whim.
And, until not long ago, a new law could impact an old one, but without explicitly replacing or quoting it.
End result: the legal authority sometimes had to recompute monthly the retiring benefits for, I think, 16,000,000 pensions, plus doing again all the actuarial computations to see the impact on the potential benefits of those who were still working.
And, of course, sometimes they had to do those computations to answer questions about the feasibility of changes.
Now, once the legal authority asked for a statistical report that wasn’t required by law- but while they were limited by law to the 10 years (I think, but I can check if your are interested) statute of limitation, there was basically no limit on the statistics that they could ask.
Therefore, when my customer, as any other large customer, provided the statistics, they called back and said: by answering to that request you showed that you are keeping data that are older than 10 years; would you mind sending us a copy of the data, as we lost the tapes?
A measure, a statistical report, implied a different behaviour- and my customer told me that, since then, they decided to keep the data forever.
Well, after few ownership changes, I do not know if their foreign owners appreciated the niceties of the legal and informal legal system in Italy, and maybe that is not the case anymore.
The point is simple: enforcing transparency isn’t enough, what matters is ensuring that that transparency is the result of a process that naturally leads to following an appropriate path.
Otherwise, you risk doing what many companies did at the beginning of the ISO9000/quality season: they kept doing everything as it was before, simply adding on each process a specific “quality” step, increasing the cost instead of increasing the quality.
Whenever I had the chance to design a KPI or measurement or dashboard system from scratch, I made a point of trying to get the approval from those who really needed the data (usually senior managers or their representatives) to focus on what they needed, and not what could be measured.
My first real cases? On a banking project, where I devised a humble spreadsheet with few measures to mark progress on a major programme of work (a banking general ledger), so that on a weekly basis I could inform the manager and partners on our status; and for a private banking unit, where in agreement with the controller I “shelved” a ponderous study with plenty of charts and measures, and designed a much simpler system based on few measures that actually monitored behaviour, and could be easily updated and used: both before 1990, and both with PCs that were less powerful than any 100 EUR smartphone.
“Big data” can deliver significant value and transparency both within the State and the private sector- but only if those technologies are used at the end of a virtous process, that usually requires a little bit more than mere number crunching.
It is called… policy setting.