Ringing the changes

With a general acceptance that police response to false alarm calls imposes substantial direct and indirect costs on police forces, alarm users, alarm companies and society at large, the pressure to act upon the problem has intensified to almost pressure cooker proportions.

The past two decades have seemingly been underpinned by a raft of measures conceived (and then imposed) on alarm system end users, installers, designers and monitoring concerns – first by central Government, then by individual police forces and, of late, the Association of Chief Police Officers (ACPO).

Punch drunk with ACPO’s regular false alarm ‘league table’ figures for the regions, the security manager could be lulled into thinking that all the necessary data on false activations needed to table a reasoned response to the problem has been formally collated, evaluated and analysed. Wrong. In reality, a lack of in-depth research has meant that policy has always been implemented in the absence of any real understanding of the topic.

There are several questions that we have failed to address, and to which answers are desperately needed. To what extent has policy worked to reduce policed false alarms? What are the likely future effects of policies now in development? What have been the adverse effects of policies developed to date, and – most important of all – can the alarms industry address the false alarms problem itself rather than continuing to submit to imposed policies?

True, there is a good deal of conventional wisdom as to the causes of false activations, much of it generated by the mountainous volumes of information dished out on the topic following the launch of ACPO’s ‘Security Systems Policy 2000’. Security managers are told that the causes are – in no particular order – “people, procedures and/or equipment” (or maybe even a combination of all three). However, the rhetoric has not been subject to anything but cursory and independent scrutiny. Until now, that is.

As National Security Inspectorate chief executive Tom Mullarkey rightly points out: “It’s the reality of bringing the false alarms problem into focus holistically, through an historical evaluation supplemented by thorough sampling, that gives our initial study with PRCI its credibility.”

The study to which Mullarkey refers is a national research and evaluation project carried out by Perpetuity Research and Consultancy International’s (PRCI) Professor Martin Gill and colleague Martin Hemming. Entitled ‘The Causes of False Alarms’ and published only last month, the 75-page report aims to assess the effects of what has been done so far to alleviate the problem, providing – as Professor Gill suggests – “a fresh look at a range of issues that have so far escaped independent attention”.

Armed with funding from the Department of Trade and Industry, Gill and Hemming have spent much time talking to alarm company staff and those working ‘at the coal face’ in Alarm Receiving Centres (ARCs). They have interviewed subscribers (about false alarms in general, and specific incidents of recent times), and then compared their perspective with that of the alarm company. They’ve also spoken to a host of systems manufacturers and engineers, looking at all false alerts – not just those relayed to the Boys in Blue.

What, though, have been the findings? First, we should examine the scale of the problem.

False alarm activations in context

According to ACPO figures released late last year, around nine out of every ten alarms that the police were called to attend in 2001 turned out to be false. At the same time, the number of remote systems approved by the police had reached nearly one million. In parallel, the latest ACPO figures state that false alarms in England and Wales also total something close to one million per annum, outnumbering genuine alarms by over nine to one.

Costing the police response to those false intruder alarms is far from easy. If officers attending such alarms are called away from what would be construed as more constructive policing, then there’s obviously an opportunity cost of their attendance. By the same equation, people like to see police out on the streets.

What we really need is a full cost-benefit analysis of all police activity if we’re to measure the indirect cost of false activations.

And what of the indirect costs to society? Professor Gill offers some headings under which these costs might be described:

l the road safety implications of police and keyholders rushing to attend premises where an intruder alarm has activated (keyholders thinking they were off duty might be unfit to drive because of alcohol consumption, and fall foul of the drink-driving laws);

l the management and staff costs of having employees attend site(s) out of normal hours;

l the time spent waiting for an engineer to attend and re-set the alarm system;

l the engineer’s time, which impacts directly on the alarm company but which might be wholly or partly passed on to all alarm users (and thence to their customers);

l further staff costs if employees who attended site out of hours are expected to attend and work normally the following day.

Studies thus far on false alarms have been few in number and – for the majority – narrow in scope. To date, the most comprehensive study undertaken was that by Fieldsend (then secretary of the ACPO Sub-Committee on Intruder Alarms) in 1994, entitled ‘Intruder Alarms: The Way Forward? Policy and Resource Implications for the Police Service’.

Fieldsend discovered that owners of systems on police response saw price as more important in choosing a system than the reputation of the various alarm companies in the market. Users off the police response list held the entirely opposite view. Users themselves cited user error as the main cause of false alarms, followed by – in descending order – faulty manufactured systems, poor equipment, poor installations and sub-standard maintenance.

“Fieldsend’s findings from his users’ survey seem intuitively correct,” states Professor Gill, “and in line with existing police and alarm company thinking.”

However, even this study conducted over a number of years didn’t find out what the data gathered and held by alarm companies and subscribers had to say about the reasons why particular false alarms had occurred.

Research methodology at PRCI

Gill and Hemming gathered information on over 15,000 alarm activations affecting no less than 2,000 subscribers over a three-year timespan. Rather than alarm companies sharing the data it was ‘pooled’ (this requires companies to allow access to their false alarm – and false alert – statistics on the basis that researchers will be able to compare each firm’s strategy on false alarm management but that no one company’s data is revealed to another).

Three alarm companies were reviewed which, for the sake of argument, we’ll call A, B and C here as is the case in PRCI’s final report (the participating NSI firms cannot be named for obvious reasons).

Company A is an independent, medium-sized alarm concern operating in the Midlands and servicing a stock of 865 police-calling intruder alarm systems. The end users it deals with cover most sectors, including commercial, domestic, retail, industrial and licensed premises. 88% of its systems are on police cover. Details were collected of 7,457 intruder alerts signalled between January 2000 and December 2002.

Of the alerts generated, 398 were identified as possibly genuine intruder activations and 426 attributed to personal attack. Thus there were 2.8 false alerts per system per annum during the study period. 89.5% of the alerts generated were false. Clearly, the introduction of ACPO 2000 and DD243:2002 has had no effect here. Indeed, the introduction of confirmation technology may have increased the proportion of false alerts. Public sector premises caused the most false alerts, followed by the commercial and industrial sectors.

Company B is a national alarm company of semi-autonomous branches. As well as installing and servicing alarm systems, the firm operates its own ARC. 55,000 systems are monitored in all. The ARC deals with around 3,400 live alerts per week, about 2,000 of which result in remote resets.

Nationally, in 2002 Company B’s intruder alarm systems signalled just under 250,000 alerts (an average of 667 per day). Almost 75% of these were filtered by the ARC, leaving over 60,000 policed activations. Almost 93% of those activations turned out to be false.

Last, Company C is also a national alarm company operating a network of semi-autonomous branches. Research was conducted solely in the Midlands area, where Company C controls 725 intruder alarm systems that are capable of being policed. Some 570 false alerts were detailed in the period 10 February 2001-17 February 2003.

“That can only be a fraction of the total number of alerts in this period,” comments Professor Gill. “We discovered in preparing for the survey that policed false alarms alone were running at 0.68 per system per year in the summer of 2002, equating to almost 1,000 policed alarms during a two-year period.”

Of the 173 alarms passed to the police in the period July to October 2002, 165 (95%) were false, equating to an annual false alarm per system rate of 0.68.

What are the causes of false alarms?

A questionnaire distributed among NSI inspection engineers showed that they are all in agreement about one factor. Namely, that system owners are the primary culprits behind false alarms. Alarm companies were next in the firing line, with ARCs little to blame and the police totally blameless. Errors on entry and exit from premises are the most common problems. Unsurprisingly, given that users are blamed the most, the next most common causes cited were premises being left insecure, accidental activation by a cleaner and the inadvertent locking-in of a customer or visitor.

Line faults were commonly cited as causes by the engineers, with non-RedCARE faults assuming a slightly higher degree of importance. Inspectors’ responses as to what single initiative would do most to reduce false alarms were varied, but there’s a general view that customers should be asked to dig deeper and either pay heavy cash penalties (if found to be negligent) or be forced to install higher specification systems.

In all three alarm companies surveyed, subscriber error was the largest cause of false alerts, accounting for around 50% of all such incidents. The peak period for false alerts was found to be between 6.00-9.00 am. Errors made by subscribers on entering or leaving their premises fall into four types: use of the wrong entry and/or exit route, deviation from the correct route, lapse of time before disarming the system and incorrect entry of the code at the keypad.

The quality and frequency of end user training was not at all high. According to Gill and Hemming, users at two of the surveyed firms didn’t feel “sufficiently connected with their alarm systems” to know what they needed. Alarm companies, then, must be proactive in offering remedial training to those members of staff who actually use the system. Upgrades cause an increase in false alarms, but also provide an excellent opportunity for refresher training.

Aside from user error, equipment malfunction and the poor siting and installation of detection devices present the largest contribution to the false alarms problem. It’s accepted in the industry that there’s a (regrettable) tendency to sell alarm equipment on the “bells and whistles” angle as opposed to ease of use. Once installed, the complexity of control equipment can – and often does – cause problems for users.

In addition, overly complex systems will lead to errors by engineers when installing equipment for users, and subsequently trying to train them to use it.

“These findings,” states the report, “offer tentative support for the proposition that the boundaries of responsibility between the manufacturer and the installation company for false alarms blamed on equipment are blurred” and that “the records of the alarm company should not be taken at face value”.