Divide And Don’t Conquer

In my last post (“Pie In The Sky”), I talked a bit about the effect of handovers within systems, and how they almost always result in waste, delays and inefficiencies. This post is going to look at this concept in a bit more detail, using a simple, real-life example from a colleague in another police force. Don’t worry if you are not a police person – the lessons from the example are so blatant that no specialist knowledge is required, and you will probably recognise similarities within your own organisation.

There seems to be a commonly-held misconception that by separating functions within a system, the organisation will somehow achieve enhanced levels of efficiency. The argument goes along the lines of, “If we create specialist teams with defined remits, they will do a better job than mere front line dogsbodies”. In a policing context, this can result in new departments being created that operate within strict parameters, along with mandatory handovers once an offender is arrested (typically by a frontline officer). In systems terms, the concept of separating out functions and establishing teams that pick up handovers from others is known as ‘division of labour’. It results in arguments about what is in whose remit, how policy is interpreted, and so on. It also slows down the flow of the work, disempowers frontline workers, and results in sub-optimal output.

In some cases however, the existence of particular specialist departments is entirely necessary and appropriate. A good police example would be the Counter Terrorist Unit, which deals in such a unique line of work that it would be unfair and ridiculous to expect a shift bobby to manage a complicated investigation around someone arrested under anti-terrorism laws. Another example would be the Economic Crime Unit, which deals exclusively with complex and high value frauds. Despite the fact that these are separate units within the overall system, the presence of such departments is not incompatible with the theory of building an effective system.

Sometimes lower-level prisoner handovers are also unavoidable. For example, officers could make an arrest 20 minutes before the end of their shift that will require extensive evidence collation ahead of interview. It would not be practical for them to stay on duty for many hours to deal with the job themselves in these circumstances. Another common factor is where the suspect is too drunk to be interviewed for several hours. It would be unfair (and costly) to expect the arresting officer to wait around for the suspect to be fit to be dealt with after he or she would normally have gone off duty.

Theoretical arguments against division of labour must always be considered against the need to achieve the stated purpose of the system (e.g. to effectively prosecute terrorists or organised criminal gangs operating international multi-million pound fraud rackets). Any notion that such top-end specialist teams are somehow incongruous with systems thinking is predicated on a rigid ideology that ignores real life. Furthermore, any interpretation of systems thinking that suggests all the workers in a system must perform an identical role is a fallacy that translates into some ill-conceived philosophy akin to ‘Systems Communism’. It is also a sure-fire way of wrecking the system.

Anyway, back on track. Division of labour and handovers are always bad when they exist unnecessarily. For example, if the first person you speak to when you report a fault with your internet service could diagnose and fix it for you, this makes so much more sense than them working through a junior-school level diagnostic flowchart before passing you onto someone else in another department because the problem wasn’t contained in their chart and it blew their mind. If the expert is at the front end it prevents handovers and results in a nice short process, as well as a faster, more efficient service. As I pointed out in the pie shop story, any handover (whether necessary or totally avoidable) introduces an additional stage into the process and therefore slows it down. The skill is in knowing your system so that you can identify any handovers that are essential, then eliminate all the others.

Enough theory for now. I’m going to run through an example featuring a dysfunctional system that thrives on separate departments with narrow remits that all receive handovers from front line workers. The example will demonstrate how this structure actually works against the system, ignores its purpose, generates waste, decreases capacity, and results in poorer service delivery. Here goes…

I have a friend who is a frontline PC on a response team in another force. We’ll call him ‘Gary’ to protect his identity. Gary’s force has recently undergone a reorganisation (or you could argue, a redisorganisation). The force has reduced the number of frontline police officers on response teams, moving some of them to various newly-generated teams with specific remits. One main feature of the new structure is that response officers no longer deal with their own prisoners following an arrest, and are obliged to hand the job over to an ‘investigation team’. The laudable aim behind this idea is that response officers are therefore not tied up in the station dealing with prisoners, and by handing the prisoner over to the investigation team, this will increase visibility, as well as availability for calls from the public. Sounds like a good idea at first, doesn’t it?

The problem is that the person who came up with this brainwave did not understand systems. There are all sorts of problems with response officers handing over prisoners as a matter of course. Aside from the horrible effects on the flow of the work that we will look at shortly, this design removes ownership and responsibility from the arresting officer. There is also the argument that it will de-skill frontline response officers, as they will become less proficient at interviewing suspects. Furthermore, smaller response teams will be less capable of managing demand at peak times (particularly in a rural force, such as where Gary is based). In addition to this there are implications for officer safety, as back-up from colleagues may be several miles away. These factors alone make this change in organisational structure a good example of introducing a systemic change that appears to offer short-term benefits, but which guarantees long-term rot.

But wait, it gets worse!

The other day, Gary told me about a straightforward incident that was unnecessarily complicated by the dysfunctional system he is obliged to operate within. It drew more people in than it needed to. It wasted time, internally and externally. It kept him off the streets. It would be funny if it wasn’t so frustrating, or costly for the taxpayer.

Gary was called to deal with a shoplifter who had been detained by staff. He had stolen low-value items and upon police arrival, he readily admitted he was guilty. (This may have had something to do with the fact that he had been seen committing the offence by three members of staff, as well as being caught on CCTV). Nevertheless, this type of incident is usually about as straightforward a job that a police officer can get. The items were returned undamaged to the shop, and the shoplifter was arrested for theft.

A check on the Police National Computer (PNC) revealed that he had not been arrested before, and the shop manager did not particularly want to go to court for such a relatively minor offence. In theory at least, providing that he subsequently admitted the offence in a taped interview, the shoplifter would usually be considered eligible for a caution. In these circumstances, this would be a proportionate outcome that would be acceptable to the shop manager. It would also negate the requirement to obtain witness statements or seize the CCTV, and mean that Gary would be back on patrol within a couple of hours at the most.

Except for the system.

What happened next was that Gary drove the shoplifter to the custody station (which was 30 minutes away, being as they were in a rural force area), and after he was booked in (this also took around 30 minutes), the following conversation ensued:

Gary: “Sarge, this one’s straightforward. He’s making a full and frank admission, so I’ll crack on and interview him myself if that’s okay?”

Custody Sergeant: “No you won’t lad. You’re not allowed to deal with your own prisoners any more. You will have to put together a handover package and then hand the case over to the investigation team”.

So, Gary had to leave his prisoner in a cell, drive the 30 minutes back to the shop, explain to the shop manager and two staff members that he would have to take statements from them after all, and that he also needed them burn off a copy of the CCTV footage. Being a quick writer, Gary was able to complete all three statements in two hours. An extra bonus was that it only took 15 minutes to burn a copy of the CCTV footage onto disc*.

(*Note: I’ve used a bit of artistic licence here: ask any bobby and they will tell you it’s always, “Nah mate, only the manager can work the CCTV and she’s on holiday / nights / at our Outer Hebrides branch this week”. Or there’s no disc available. Or despite fiddling with it for hours no one can get the damn thing to work. Or once you get the disc back to the police station there’s no machine that recognises the format. Etc. Etc.)

Once armed with all this evidence, Gary drove the 30 minutes back to the police station, where he then commenced tackling the gargantuan bureaucracy of the infamous ‘handover package’. This varies from force-to-force, but as a minimum will require the arresting officer’s statement, statements from all witnesses, a narrative of the incident from start to finish, as well as printouts of all relevant documents relating to the case, including a copy of the incident log, crime report and associated papers, a copy of the officer’s pocket book entry, PNC print, exhibit details (i.e. CCTV disc), along with a printout from the detained property computer system once the exhibit has been booked in. Along with this, there is almost always a pro-forma that lists all the papers attached to the handover package, which also requires a supervisor’s signature.

Needless to say, compiling the handover package to this standard is not a quick task, although Gary managed to do it in just an hour and 15 minutes. He then took it to the investigation team supervisor, who had to sit down and check through it, before endorsing the papers that it had met the required standard for his team to take on. Despite the likelihood that the outcome in this case was probably going to be a caution (meaning that the statements, CCTV evidence, and other documentation would never be required), policy dictated that all these elements must be present before the investigation team would deal with the shoplifter.

At this point, Gary was able to go back out on patrol. This was six hours after the ‘straightforward’ arrest.

The investigation team supervisor then had to find and allocate one of his officers to deal with the case from that point onwards. Of course, this officer had absolutely no knowledge of the matter whatsoever, so he had to read all the handover papers, including the three witness statements and Gary’s arrest statement, as well as view the CCTV. This prepared him for the interview with the suspect, which commenced about an hour after Gary exited stage left. The interview and subsequent administration of the inevitable caution took just an hour, before the stopwatch on the whole process ended 8 hours after the arrest.

The end-to-end process for what happened is mapped below:

Custody Process Flowchart 1

By following this model the following consequences have been caused:

• Gary was off the streets for 6 hours.
• Three members of the public have had to provide statements that weren’t needed. (Their time is not free!)
• A whole host of unnecessary paperwork and preparatory activity was generated to meet the standard to pass the job across to the investigation team.
• More officers were drawn in to the process than was necessary. (Their time is not without cost either).
• The end-to-end time was unnecessarily extended.

Contrast this with what the process would have looked like had Gary been allowed to retain ownership of the case:

Custody Process Flowchart 2

The advantages are obvious:

• No handovers.
• Ownership retained throughout.
• More efficient end-to-end process.
• Gary is back on patrol within 2 hours.

Considering that Process 1 was intended to increase the time response officers spend on the street, I hope you share Gary’s sense of frustration that this poorly-design system actually achieves the opposite. It also ‘achieves’ the introduction of blockages in the cell block, as the suspect is in custody for 8 hours instead of 2, meaning their capacity to manage other prisoners is affected. It causes a backlog in the control room as officers like Gary are tied up with these sorts of incidents four times longer than they should be, meaning there are less officers (on an already reduced response team) left to deal with incoming incidents. This means that outstanding jobs stack up and the public receive a sub-optimal response. All because someone with good intentions interfered with a system without understanding it. Ironic isn’t it?

The argument is not against every handover or specialism, but against rigid organisational structures and inflexible policies that dictate that there will be handovers and division of labour regardless of circumstances or context.  

The lesson here is simple:

The more an organisation compartmentalises its functions, the more handovers are introduced. The more handovers that are introduced, the more waste is driven into the system, and the less efficient it becomes. As the system becomes less efficient, service delivery gets worse.

The solution is even simpler:

Remove unnecessary handovers.

I thank you.

(Note: ‘Gary’ was not paid for contributing his story, and no animals were harmed during the making of this blog).