The Give and Take of Information

veryard projects > information management > give and take

we offer

about this page

on this page

other pages

education, training and skills transfer

model reviews

information and advice on methods and tools

supporting materials - including patterns

This paper explores some political issues of data architecture and data processing. We link these to issues of business strategy and strategy ceiling, as analysed by Philip Boxer.

The Screen

Visibility

Technological Service Focus

The Ceiling

Tactics

Data Stratification

Thinking the Unthinkable

The Screen

veryard projects > information > give and take > screen

The computer offers information as services through a screen. The screen is both literal and metaphorical. It is a surface on which the data are presented, and also a filter that controls what the user sees. screen

As with any technology, a computerized information system alters what is visible (which objects attract your attention) and what is invisible (which objects are taken for granted).

The first (obvious) level at which this is apparent is in the distinction between source data (sometimes called raw data or atomic data) and presentable data (which I elsewhere called ‘cooked’ data, in an ironic reference to Levi-Strauss). Lots of MIS processing is devoted to converting source data into a form that MIS professionals deem more meaningful (and therefore presentable) to ‘users’. Among other things, this typically involves

Filtering out ‘bad’ (i.e. erroneous or incomplete) data
Use of simple statistics to reduce the volume of data (especially subtotals, averages and ratios)
Conversion of data to a single homogeneous format, supposedly to aid ‘comparison’ between data

But this distinction is relatively open. MIS professionals appeal to the concept of ‘drill-down’, whereby ‘users’ are able to retrieve the raw data on demand. This apparent openness further obscures more subtle invisibilities.

WIGO - What Is Going On

Visibility

veryard projects > information > give and take > visibility

In our view, business strategy is not a fixed outcome of an annual management workshop, but an ongoing thinking process. The strategy process creates a demand for data. To express this more fully: we postulate a manager engaged in strategic thinking. (This manager may be a person or a team, or some other entity playing a managerial role; for the sake of simplicity let’s talk of the manager as a simple person.) The strategic thinking process exposes some lack which is interpreted as a need for data. The manager therefore positions himself as a user of data, and makes demands for data.

At its extreme, the demand is stated as follows. We need such-and-such data in order to be able to make a sensible choice between the strategic options available to us - or even to determine what the strategic options are.

The data user then tries to perceive (make out) a data object that fits this demand. Sometimes, the user is persuaded that the computer system provides the requisite data. In other words, that the demand is already satisfied by the computer system. (The ceiling is so low, that any mismatch is already above the ceiling.)

But sometimes the MIS professionals fail to persuade the user of this match.

Sometimes the user believes that the data exist behind the screen, but are being denied to him by the computer system and/or the MIS professionals. (‘I know the data are in there, but I can’t get them out.’) Perhaps we may call this frustration.

Technological Service Focus

veryard projects > information > give and take > technological service focus

From a technological point of view, computer data systems are believed to be complete, relative to some defined scope. From this point of view, there are really only two things you can do to improve the data services:

Improve content: extend the scope, provide access to new data objects

new sources of data
data reprocessing (e.g. statistical or combinatorial analyses, to generate more complex data objects)

Improve availability: provide more, easier and more reliable access to data

better quality information (more timely, relevant,

whenever the user wants
wherever the user wants
more ‘user-friendly’
more cost-effective

A technocrat will usually be willing to undertake such improvements, provided the quality-of-service criteria remain unquestioned: what is to count as timeliness, relevance, accuracy, completeness, consistency, user-friendliness and cost-effectiveness. Who, for whom and why are above the ceiling.

The ceiling

veryard projects > information > give and take > ceiling

The data structure represents the organization’s power structure. In many organizations, this is "above the ceiling". In other words, it is not accessible to rational management.

For example, a given accounting system reflects a given way of distributing financial value through the organization. Thus a shift from (top-down) allocation costing to (bottom-up) activity-based costing can be a highly political move. In many organizations, such a move cannot be properly discussed, because it is above the ceiling.

Data ownership is also a political issue. Who in the organization gets to define what counts as a customer, or what counts as a sales order? This issue is perceived by technologists (naturally) as a technological issue, called data sharing: how can we get all the users to use the same customer and sales order data? (In the 1980s, the technical solution (S2) was a central database, in the 1990s, the technical solution is a client-server architecture.)

The converse can be useful. We can often take a lack of data sharing as a symptom of the power structure. This lack (a) is keenly felt by data technicians, who regard it as prime evidence of inconsistency ($). Data sharing, and the particular form of consistency demanded by data sharing, are Good Things (S1).

Tactics

veryard projects > information > give and take > tactics

Philip Boxer identifies two ways to move the ceiling. He argues that a sideways push can often be more effective than a direct upwards push.

Push Upwards	Push Sideways
Move the ceiling by directly attacking the power structures that are reflected in the data architecture and the associated strategy process.	Challenge the fixed positions inherent in the data architecture and the associated software process.

So how do we intervene in an organization’s data architecture? I suspect we need simultaneous guerrilla attacks on a number of fronts:

Attack the computer data architecture (structure and process), with particular attention to the lack of integration and ‘consistency’. Expose (to the MIS professionals) the business ‘realities’ underlying these lacks. Challenge the expectation that new technologies (such as client-server or OO) will address these lacks.
Attack the software process, with particular attention to Requirements and Quality. The SEI Capability Maturity Model encourages a positional software strategy at levels 2 and 3, but moves towards a relational software strategy at levels 4 and 5. If the software organization is still at level 1, this could be a very long-term campaign.
Attack the accounting systems that attach (financial) data values to business objects.
From the user side, formulate (complex) data objects that cannot be constructed from the existing data systems.

To sum up, we are looking for various forms of lack or impossibility - in the user organization, in the data architecture, in the software process and technology - and trying to link them together, so that the Real cannot be localized in any one of these.

Data stratification

veryard projects > information > give and take > data stratification

This preliminary exploration encourages me in the belief that Philip Boxer's Pan structure can be applied to a stratification of the data architecture.

For business read data. For market read data usage.

veryard projects > data > give and take