Home » Uncategorized

Operational Data and Social Justice

I spotted an interesting book in my local library recently:  The Final Report of the Truth and Reconciliation Commission of Canada [1].  I thought to myself, our government spent considerable resources on this commission.  I should at least browse through the final report.  I flipped through the first few pages.  I found a note saying that the contents are public domain.  In this blog, I reproduce some of the contents of the report to create a setting for my discussion on operational data.  There are different ways to interpret the role of the commission and its recommendations.  I will be discussing how operations – in a business or institution – can become disconnected from social justice due in part to how data is handled.  (I am tempted to add “environmental stewardship” to “social justice” – but here I lump everything together.)  I suggest that this dualism between operations and social consequences threatens organizations not merely in relation to politics and public relations but also market adaptation and structural transformation.

Canada is responsible for – and I would say slowly recovering from – cultural genocide.  We systematically took actions to purge the country of indigenous culture.  Although I wasn’t around at the time, I nonetheless feel the weight of this injustice.  I contextualize it like most people: e.g. “What if this had happened to me and my family?”  Then interestingly enough I bring it into the context of the workplace: “Is there anything I can do to ensure this doesn’t happen again” and “Had I been around at the time, could I have done something to prevent this hate crime?”  I deal with so much data every day.  I recognize that facts and figures can sometimes become a sea of silence – and that expression can be rendered void of voice.  This is a social justice question.  It is also an important business consideration – how data intended to help an organization understand important events might instead do just the opposite.

Not that I have spent a great deal of time studying them, but I am under the impression that the Nazis kept fairly good records of items taken from prisoners, who themselves were coded or numbered for identification.  Irrespective of the horrors they inflicted on their prisoners, I have no doubt that Nazis had comprehensive systems of “due process.”  The successful execution of these processes does not diminish the underlying social injustice.  If we turn our gaze away from the Nazis specifically and focused our attention on the dualism between process and consequence, I suggest that certain elements can be found in more contemporary examples:  e.g. manufacturing of air bags and injuries from their malfunction in vehicles; administration over disaster relief and lack of tangible improvement; allocation of funds for indigenous program and chronic levels of suffering at the community level.  It is like the data serves a master other than those meant to receive the service; and the compilation of ever larger amounts of data does little to help members of society.

Operational Data and Social Justice

The image above (from the report) shows a ceremony called the “Sun Dance”:  it was a source of indigenous spiritual identity that would eventually be suppressed [2].  To this day, it is unclear to me how such prohibitions could have been legitimized.  I said that this blog is about “operations.”  The commission examined a system of institutions involved in taking children away from their parents.  The objective was to sever these children from their cultural identities.  I will be considering data as it relates to the operation of this type of institution on an abstract level.  Consider the quote below from Sir John A. Macdonald, Canada’s first prime minister, explaining the rationale behind the “residential school system.”

When the school is on the reserve the child lives with its parents, who are savages; he is surrounded by savages, and though he may learn to read and write his habits, and training and mode of thought are Indian.  He is simply a savage who can read and write.  It has been strongly pressed on myself, as the head of the Department, that Indian children should be withdrawn as much as possible from parental influence, and the only way to do that would be to put them in central training industrial schools where they will acquire the habits and modes of thought of the white man. [3]

I don’t believe any residuals of these residential schools remain in operation.  But consider the operational environment of such a place if it existed today.  In my day-to-day routine, I keep track of the work that people do.  I check for patterns and trends.  I monitor not just what they do but what “they should have done.”  I collect operational data particularly as it relates to performance.  In light of these tasks, I find myself wondering what kind of framework would be needed to bring “wrongness” to light in an operational setting.  How could the pernicious levels of injustice in a residential school system persist in a setting where all sorts of data is collected and analyzed by many people – including professionals?  In Canada, Sears is currently in creditor protection.  I am left questioning how a company that had the opportunity to interact with customers every day somehow floated away from their needs and interests.  In a literal sense, the company “capitalized on its failure” – that is to say, it used its capital to sustain a failing structure.  This is possible on the duality between operations and consequences.

The next image shows children in a residential school doing what the report describes as “institutionalized child labour” [4].  It seems children were not merely losing their identities in these schools; in some cases they were compelled to work.  Some were injured while working.

Operational Data and Social Justice

A problem with quantitative data is its poor ability to express the structural details behind the numbers.  The child labour case depicted above might find itself quantified in terms of production, hours of labour, and incidents of workplace injury.  The validity of this quantitative expression does not negate the social injustice of the underlying situation, which I will describe here as child slavery.  This is the term that should be used when children do labour in a setting where participation is required.  Apart from slavery, the obvious criminal activity that comes to mind is “forced confinement.”  The processes that occurred in these residential schools weren’t about helping children acquire the “habits and modes of thought of the white man” per se – but rather to utterly realize his victimization.  I want to emphasize how the use of metrics in such an environment – without inclusion of details about the setting itself – represents and absolute and extreme extension of colonialism that persists to this day in organizations.  The toxicity of an environment is not merely a physical and chemical phenomenon but also social.

The next image shows children in a residential school sick from diseases such as tuberculosis [5].  I ask myself if it might be possible to incorporate ethics and social indicators within the operational framework – e.g. in the context of data collection.  The use of metrics without structural details invites analytic disassociation and administrative abuse.  I have found that it is possible to maintain a “parallel event stream” to coincide with metrics.  Structural details are problematic in that they are sometimes difficult to quantify – or they might be non-quantifiable.  There might not be an effective way to express event data using traditional statistical methods or conventional computer applications.  However, event data need not be quantified or quantifiable in order to be part of a data-collection regime.

Operational Data and Social Justice

Ownership – Power to the People

Important when considering the use of data is “ownership.”  I don’t mean who owns the data but who brought about its meaning.  Quantitative data as it is collected in an institution belongs to that institution in a fundamental way.  What outsiders can obtain is determined by what insiders wish to give.  Even those working within the institution cannot stray beyond these design confines – i.e. to express more than what the system is designed to give.  In contrast, event data can be formed by the narrative of the individual.  There is no need for the data to fit into a prescriptive quantitative regime before gaining relevance.  Without a parallel event stream and methods to make use of this type of data, the institutionally-mandated constructs would have little context and meaning for outsiders at least in operational terms.  There would certainly be budgets, audits, and financial reports.  I point out however that corporate scams sometimes involve budgets, audits, and financial reports.  We have merely the illusions of legitimacy.

There is no need for the institution to prescriptively deliver a systematized information system meant to exclusively convey metrics important to managers.  Those intended to be served by the system know what is important to them.  There is no need to prejudge data arising from their sentiments, thoughts, and emotions – to pigeonhole these details into a metrics regime.  The fact that their data might not fit the prescribed normative is hardly relevant in the context of a parallel event stream.  Nor does the parallel stream necessarily interfere with the administration of the institution.  Similarly, it isn’t necessary for the clients of an organization to conform to the interactive parameters afforded to them, in the manner prescribed, using constructs conceived during business meetings.

Today in Canada, we are measuring the outcomes and consequences of the residential school system.  Like the school system itself, the methods of response have likewise been institutional.  In relation to the response, I would suggest there has been a failure not so much of the “institutional response” but of the “system of institutional response.”  It is a system that has difficultly going beyond the metrics to effect structural changes in communities.  When the metrics lack structural details, I would expect misallocation of resources, inability to adapt, and problems guiding policies.  It has recently been suggested by some that certain problems relating to indigenous communities might be a question of decentralization:  i.e. high suicides levels among indigenous youths are due to administrative and decision-making dysfunction.  However, I would say the obstacle is more primal:  it is quantitative alienation – brought about by the colonization of data.

Parallel Event Stream


Operational Data and Social Justice

In my use of parallel event streams, the events are tokens or symbols normally associated with metrics as in the example above:  events from three departments (departments A, B, and C) are associated with the closing price of a stock for the company.  The general idea is that certain departmental activities might contribute in some way to the closing price.  (The assertion is not necessarily true.)  I recognize that there are important considerations such as timing and the decay of influence – none of which I can cover in this blog, unfortunately.  On 2015-02-19, department A brought about a stock split as per the symbol “[999]”; this explains the sudden decline in price.  Instead of price, it is possible to use units sold.  Instead of departmental events, competitor or market events might be used.  Indeed, there need not be a clear or coherent relationship between the metrics and coincidental events.  That having been said, it is reasonable to draw coherent relationships in the analysis: e.g. suicide attempts in relation to personal, family, and community events in parallel.

Probably the most straightforward way to synchronize quantitative and event data – that is to say, to make them “parallel” – is chronologically, as in my example.  Production metrics gathered on a particular date could be associated with events related to operations at a production facility.  Units produced could be connected to a string of quality concerns (events).  These linkages are not hard as one might expect in a defined data object but rather quite soft; for example, the discursive events of a particular employee could be attached to production metrics – although he or she might not be directly involved in production.  The parallel association is therefore not meant to be interpreted as causal but coincidental.  Administrators can bundle into the parallel event stream any events that seem relevant while the issue of inclusion or exclusion can be assessed during more focused analyses.  Synchronization need not be chronological.  It can be spatial or geographic.  Events can be distributed for instance in relation to locations in an institutional facility – e.g. cases of tuberculosis or child molestation.

Below are some parallel events from my personal database – dated last Christmas.  Notice that I use fairly readable English for my event codes rather than code reference numbers.  I don’t know if readers have ever been asked questions like, “Can you explain the sudden increase . . .” or “Why did this decline so sharply . . .” or “Can we explain the difference that occurred between . . . and . . .”  These are the sorts of questions that might be best addressed using parallel events rather than metrics alone.  Metrics are ideal for monitoring outcomes, but they poorly designed for studying incomes.  When determining how to deal with an outcome, it is necessary to create policy in relation to the income.  Without analyses of parallel events, funding is likely to be used only to “look into” problems; or the funds might be ineffectively allocated.  From a research standpoint, parallel events help bring to light the circumstances behind the outcomes.


















Tokenization and the Schematics of Ontology

The use of event symbols is about giving reality a means of expression – but not effecting it necessarily for the sake of mathematical analysis.  The process of handling events in relation to metrics is something I have been writing about for many years.  I use a system to examine events in relation to metrics called the “crosswave differential algorithm.”  In this blog, I merely focus on the logistics of recognition – the schematics of ontology.  In an information system, these schematics point to an exercise in conformity and standardization.  However, we should not have preconceptions that require agreement or impose conformity on matters that are not fully understood.  The objective should be on accommodation rather than conformity.  Data collection in an operational setting is not meant to be a race leading to a finish-line, supporting or refuting particular hypotheses.

I remember a number of years ago coming across a young girl at a bridge, clinging to the base of a lamppost.  Her wrists were slit although the blood had dried.  I suspect that the idea was for her to lose consciousness and fall into traffic.  The fall would have been enough to kill her – the bridge being maybe 150 feet off the ground.  At some point while waiting for the police to arrive, I decided to stand next to this person.  On close examination, she seemed less than 12 years old and likely indigenous.  Although I was close enough to hear her breathing, I didn’t have anything to say.  She was having a bad day; and nothing that I had to say would change matters.  I thought how difficult it is to communicate in a constructive manner to a person who is caught in storm.  I think that an administrator for a residential school probably dealt with these types of situations.  Ideally it would be desirable to anticipate and proactively engage problems before they escalate.  However, the recognition of phenomena should not be imposed.  Data is best left as a matter of self-construction.  Because if I were to try to “fit” or “force” such a girl’s situation into an institution’s box, I might be contributing to her victimization – e.g. by silencing her – by predefining her situation and imposing its place in the scheme of things.

I reflect back on that day at the bridge when I think about generating data from the recognition of events.  A parallel event stream is not necessarily any more useful than its coincidental metrics if the context of the events is controlled or has been commandeered by external interests.  I suggest however that control in an organizational context is more a matter of social construction than deliberation.  Sometimes even well-intentioned organizations will do things that are destructive merely by the imposition of control over the schematics of ontology.  This is why I say that institutions fail not just by their response but by their institutional response system.  Within operations, there is a need both to administer the intended service while nonetheless being sensitive to how those services are received and the needs of the service recipients.  The job hasn’t been fulfilled merely by its execution.  To structurally insulate the organization from the realities of social injustice is to make administration a tool of colonialism.  We have a role to sensitize the data system to the truth; this can be done through the use of parallel event streams and by enabling the expression of events of internal narrative.

[1] Truth and Reconciliation Commission of Canada, Final Report of the Truth and Reconciliation Commission of Canada (Toronto, Ontario: James Lorimer & Company Ltd., 2015).

[2] ibid., 54.

[3] ibid., 2.

[4] ibid., 78.

[5] ibid., 94.