Positive Feedback and targets

Positive feedback in economics explains why some products take off and become the standard, while others don’t.  One example often cited is VHS, which edged it over Betamax (a product with a technical edge) by virtue of having a slight lead in the market.  Because more people had VHS, more VHS cassettes were produced with more choice, making that the choice for those purchasing a system, so more VHS was bought, and so forth.

Another example is a new service being provided by a national charity.  Many local authority commissioners have quickly bought into this service, although it is not yet out of the pilot stage.  I wonder if this is the case of  “everyone else has one, so it must be good”?

Apply this to behaviour at work as well.  If someone gets positive feedback, they believe something worked and will do more of it.  On the other hand, negative feedback, fault finding etc. only tells them what not to do and can lead to risk aversion.

Trouble starts here, as positive feedback can lead to poor decisions – technically we should have ended up with a Betamax standard.  And who remembers when bell-bottomed trousers were the rage (and everyone had them)?

To encourage positive behaviour at work, people set targets.  This would be fine if we knew what would be effective, but the specialised nature of modern work means that the impact of our actions on the whole enterprise is unclear.  Targets are best guesses, which then lock people into behaviour patterns that may be damaging, or parochial and narrow in outlook.  Missed targets can turn potential positive feedback to negative or blaming.

Positive feedback appears to drive a lot of behaviour.  Targets drive a lot of behaviour at work.  How can we know it is effective behaviour until long after the fact, if ever?

Emergent learning, targets and support

I developed the model above about a year ago, but last week found a practical application for it.  You may notice I adapted it from Mintzberg and Waters model for a planned vs. an emergent strategy, with Nonaka and Takeuchi’s idea of combination thrown in.

I was discussing ways of engaging with homeless and vulnerable clients, in a workshop for managers of a not for profit organisation.  The model helped us think about how planned targets (expressed in the “Supporting People” world as “outcomes”) were very difficult to plan for.

The process is expected to run like this:

Client meets professional and an assessment is carried out.  The client expresses aspirations and they set goals together.  Over a period of time those goals are achieved through a series of small steps, and the outcomes desired are achieved.

What often happens is this:

Client meets professional and an assessment is carried out.  The client expresses aspirations and they set goals together.   At the next meeting the client has done nothing towards these and instead expresses other aspirations.  This happens a few times, the professional expresses frustration and begins to nag the client to do stuff.  The client begins to feel bad and votes with their feet.  Personal change is even further off for the client.

This happens for two reasons.  First, the insistence on setting targets and having a plan early on, before the client has explored their current situation enough.  Secondly, the professionals involved are normally generalists, not trained therapists or coaches, and they have targets to meet, so their urgency is to get action and results at the expense of exploration.

Back to the model.  If you realise that a lot of value comes from what emerges unforseen, you may be less likely to grab the first plan that comes along and try to stick to it.  Where achieving targets and outcomes are concerned, this suggests that the right target can only be identified late on in the process and that trying to nail it early may be counter productive.

Star Struck – outcomes measurement, distance travelled, and proving the value of support

I was reading a back issue (May 2008 ) of the SITRA bulletin, where I found an article on the latest metamorphosis of the Outcomes Star developed by Triangle Consulting.  This was the mental health version, the Recovery Star.

The Outcomes Star has been used in the Homelessness field for some time, and has gone through various versions.  All of them measure, usually on a scale of 1-10, a ten dimensions that represent areas of a client’s life.  The Recovery Star includes Work, Relationships, Addictions, Trust and Hope and similar.  The original measured different areas, such as offending, managing money and meaningful use of time.

This is a fine shot at measuring what is often thought of as unmeasurable, and it has become popular with several large homeless charities, some local authorities and been adapted as an idea by others, including to measure needs according to the Supporting People Outcomes areas. One point of doing this is to evidence the work done in supported housing by capturing “softer” outcomes, to justify grants and government expenditure.  Another is to provide a basis to work with clients to help them set targets and record the achievements they make.

However, I’d like to express some reservations about the basis of the underlying idea.

First, the model is based on identifying deficiencies, or needs.  What about a person’s strengths, assets and aspirations?  These can be worked in, but the need comes first.  To my mind this continues an old voluntary sector theme of seeing clients as a bundle of needs to be helped, creating dependency. It also provides labels that can be self fulfilling – if a client thinks they rate poorly in an area then it is likely they will act in that way.  Also, from any kind of performance measurement perspective it can be a nightmare, as ratings will go up and down depending not on progress, but on how much is disclosed or known at the time of a particular assessment.

Second, what you measure is what you get.  Although dressed up as a star, essentially it is a series of disconnected linear measurements.  While a skilled worker will help a service user see the links between the different areas, and be aware that going forwards can contain many dead ends and needs to go back, the model does not promote this.  As this is what we are measuring the areas identified will be what we get.  The danger is that the complexity of individual behaviour is reduced to fixing particular areas specified not by the service users themselves, but by the monitoring system.

I would argue for alternatives.  Brief Therapy ideas provide an alternative as do most coaching models.  Starting with aspirations and what a client can do, rather than on the past and needs, can form a more positive way of moving forwards.  Otherwise clients may stay stuck in the past, may pay more attention to what is stopping them than what they can do, and not achieve the outcomes that we so desperately want to measure.