top of page

Data insights

Some snippets of interesting insights that we've recently observed

Working at home

The cohort of 1989

Across crime types such as violence and stalking, we've been noticing in police areas in different parts of the country that those born in 1989 and the surrounding years are consistently showing the greatest propensity. Similarly this group are more likely to be involved in car accidents and be unvaccinated for Covid19. We are still exploring what is driving this, what it means for services and what can be done. 

Violent crime and whether its the weather

When the weather gets warmer we are finding that violent crime increases. But the effects are quite specific to certain types of violence, in certain types of area and with certain types of offender. Not only that the day of the week matters, as does recent temperature trends, while for some, its cold weather that drives up activity. The effect was much stronger than other proposed drivers of violence such as football meaning its more important to check the forecast than the fixture list.  This knowledge helps to predict demand and plan better for provision of services.

How to break your process

In a contact centre, reminder letters were often sent on a Thursday.  This meant responses were typically received the following Monday. But because Mondays see the highest average customer requests for new work this was tending to amplify an already high demand signal. Detailed simulation modelling of the process showed a counter-intuitive result. Delaying reminders to the Friday had the surprising effect of improving performance because it made the overall flow better. Details will differ in each circumstance but we learned that sometimes, process can get unintentionally 'sabotaged' by well intentioned assumptions.

Keeping the flow

In a police control room there were issues with performance. The target was to answer 999 calls within 10 seconds. However the maths of queues means this is unachievable unless there is substantial surplus capacity for much of the time. The time from when a call is made until an officer was dispatched to the emergency averaged over 4 minutes.  The analysis made for a convincing argument that relaxing the call answering time to 20 seconds freed up enough resources for use elsewhere in the system to reduce dispatch time to just over 3 minutes, improving end to end service time at lower cost.

The power of leadership on productivity

We've been observing in various contexts how leadership is like a magic bullet for productivity. Investigating numerous process and system improvement opportunities has shown that the best performers are achieving up to twice that of some colleagues doing equivalent work, and the difference comes down to leadership.

Learning from existing behaviour

Maybe leopards don't change their spots but we're finding that one of the depressingly best predictors of future criminal activity is past behaviour. But not always. Sometimes the best predictors are when, not if, offending occurs and how.

How to stop a criminal

Arrest, charge, caution or restorative justice?

Detailed review of future offending shows that 'it depends'. For example, we're finding big differences in what works for youths involved in minor crime to approaches for dealing with sex offenders.

Selecting the right projects is hard!

When it comes to selecting the right projects things get difficult quickly.  In one case, an organisation had hundreds of projects to prioritise but only a finite amount of time and resources. By using leading edge analytical techniques we were able to find a sequence and selection that gave better results at 28% lower costs. 

We're finding that in typical portfolios of projects 10-15% more value can be extracted by automating the prioritisation process, as well as giving it the science to reduce the influence and pressures that distort effective prioritisation.

Stereotypes of domestic violence are changing

A decreasing proportion of domestic violence incidents reported to police involves the 'typical' model of a male offender and a female victim who are in an intimate partner relationship. This now makes up approximately half of domestic violence reports with the remainder involving other family members, same sex couples or male victims and the trend in reporting of other forms of domestic violence is accelerating.

Deprivation and violence

In some of the most deprived areas 1 in 8 males aged 30-35 have been linked to a domestic violence incident. This depressing statistic is vital for partnerships to ponder and fit solutions to the context.

bottom of page