13 June 2012

SLAve to convention

Conventional FM wisdom tells us that we should manage our client/supplier relationships with SLAs, but maybe some of us would be better off ditching them altogether.

Service Level Agreements (SLAs for short) are an agreement between the customer and supplier about the extent, the quality or the speed with which a service is provided; 90% callout within 4-hours is a commonly used example.

In theory, SLAs are a way for the customer to make it clear to the supplier what standards they expect (that’s the Service Level part) and for the supplier to confirm that those standards can be delivered (that’s the Agreement part).  Measuring how well the SLA is achieved enables both parties to be clear about whether the service being provided is meeting expectations; this in turn helps the supplier to prioritise resources and actions to help ensure the customer gets what they wanted. That’s the theory anyway.

The trouble is; I don’t really buy into it. 

Stick to carrots

One of the big issues I have with SLAs is that they are often written into service agreements; effectively they become glorified “get out” clauses, to enable customers to terminate suppliers who fail to meet the mark.

Of course I’m not saying that having the flexibility to break of relations with a supplier who is seriously underperforming is a bad thing, but I wonder if the approach is a little short-sighted: human nature dictates that if you are going to give someone a stick to beat you with, you make that stick as small as possible. 

Does the use of SLAs actually encourage us to strive for mediocrity rather than success?  Remember that agreement part of the SLA?  Well it often seems to me that suppliers are reluctant to agree to SLAs that they couldn’t easily meet. For example, let’s go back to the 90% callout within 4-hours example cited at the beginning of this blog.  I work in Berkshire. I’d happily place a wager on the fact that in 4 hours, I could be in Swansea, Manchester, or maybe even France, but the SLA would have me believe that just getting to my office from their's (which probably isn't in Manchester by the way) in the same timeframe is some kind of indicator of successful service provision.........  So does that really drive performance in the way I’d like it to?

Not that suppliers get all the blame here.  Realistically, the service provider could probably manage 90% callout in 2-hours, but if the price of failure is losing the contract, you can’t really be suprised if they choose to play it safe. Perhaps the better approach is to prefer the carrot to the stick.  I suspect we prefered incentives to penalties, then service providers would be more willing to accept  challenging targets and this would drive better service. 

Sample size

One of the other concerns I have with SLAs is sample size.  SLA achievement data is supposed to be representative of the service performance, but for statistical data to be useful, it needs to come from a decent sized sample.

My career to date has always been in client-side (in-house) FM; I’m generally a buyer, rather than provider, of services. The supplier relationships I’ve had have been small in comparison to say, for example, large TFM relationships. The truth is, that in my area of operations, many of the standard issue SLAs fall down as a result of scale.

Let’s say, for example, my vending contractor offers me that 90% callout in 4-hours SLA.  Each month they report 100% success and everyone is happy because they get here so quickly?  Well not quite.

You see the reality is that the machines are quite new. The 100% success rate is actually telling me that they rarely break down, not that the supplier responds quickly. If the machines only break down once a month or so and the SLA is reported every month, there’s really only two ways it can go: 0% success or 100% success. Simply put, the volume of transactions is just too low for the SLA to be a useful indicator of success. Perhaps reporting it over a longer period (say a year) might be better, but I suspect that it would only really work over a much larger contract.

The other thing I really need to ask myself is, if they always get here in 4-hours anyway, should I really be focussing my energies on an element of the service that has room for improvement?

Choose wisely

Despite my opening statement, I’m not actually suggesting we all dispense with our SLAs, but I do think that both customers and suppliers need to give much more careful consideration to what their SLAs are doing for them if they want them to be effective.  Measure what is important to the relationship: use SLAs to drive positive success by measure the areas you want to improve. Don’t just measure what everyone else is measuring; don’t just be a SLAve to convention.


  1. Good interesting article Jason.

    I agree with what you say about SLAs and am often dissapointed when I hear of service providers saying they have got there within 4 hour and have therefore achieved their SLA.

    I have experienced this a few times, only to have the engineer or whoever is fixing the issue to spend hours sorting the issue out. I think Fix times are much more important than response times and provide more accurate information about the quality and speed of the service..

  2. Thanks for posting some comments Will. You're absolutley spot on. We need to measure our suppliers on the things that are actually important to us and where we want to drive performance, rather than just taking out-of-the-box SLAs because the supplier provides these to everyone.

  3. Great article as always, Jason.

    Completely agree with the carrot/stick example. So often these days we hear about the benefits of using the carrot approach to motivate and incentivise (apologies for the jargon!) our own in-house staff, so why not use the same approach with contractors?

    1. Incentivise not penalise - it's bold, but don't we always ask our suppliers to bring us innovation? Surely doing something different and better is innovation in action?

      Thanks Chris, appreciate the comments.

  4. Well put Jason.

    The problem isn't the principle of having SLA's or in having KPI's to measure them. The problem is too much focus on input and output and not enough on outcome.

    Who cares how quickly they get there, or how long they take to fix the kit? What matters is the impact of not having your machines working.If this disrupts business then by how much? If their absence is benign then why worry about it?

    When we focus on our contribution to the organisation's success instead of counting our building blocks we start developing really meaningful SLA's

  5. Thanks Martin, it's great to have feedback from an FM legend such as yourself and encouraging to know our thoughts head in similar lines........... I certainly think FMs need to focus on the right things if they are going to demonstrate value to their customers in turn. After all, do the FM's customer's care about response speed, or is it total down-time that concerns them as per Will's comments?

  6. HI Jason and all
    Incentivise not penalise, couldn’t agree more,
    Sla’s / KPI’S for in house staff? I work in the public sector where salaries are very structured with very little room to incentivise people who are willing to go the extra mile and adversely those who perhaps don’t try quite as hard as they could?
    CPD is a must for all employees at all levels, we also sometimes overlook the personal development of our staff as well as the professional and they aren’t necessary linked.
    SLA’S etc for contractors? undecided as to the extent their worth as you stated contractors are unwilling to sign up to anything to onerous and therefor they can become a little “loose” but that said they do provide a base line, a minimum level of acceptable service. As against them being a bench mark for exceptional service. They do provide a base and a format for discussion in relation to performance and service level and can be useful if things aren’t going quite so well.
    My opinion, they are or should be more useful than just a tick box exercise. If they are not then either the SLA’s or the persons involved in the process need reviewing to ensure the organisation does reap the benefits

    1. Thanks for the comments Robert. It's absolutely true that we get the best out of anyone, suppliers employees, etc, when we support a culture where it's ok to make mistakes but sucess is rewarded. I've heard it said many times, that if you don't get things wrong from time-to-time, you're not taking any chances: the same can be true with suppliers.

  7. Bonjour Jason et al...

    SLA's are great for setting expectations, "we will arrive on site within 4 hrs of a callout being logged" that gives peace of mind to the buyer of the service.
    And the message can be relayed to the end user, so they have a rough idea, when their problem will be looked at.
    that is total quality management

    I've never (apart from 1, 2 month occasion)been involved with a contract that had financial penalties built into it, so very often even though these callout response times were set in the SLA, the supplier would not meet them because of other business commitments, therefore maybe that is the time for a financial penalty and or pay for a premium service, dependent on the importance of the kit and faults to be fixed.
    So one could argue that the SLA is used as a big stick (not literally)by the customer to make sure the supplier delivers the service that is expected and needed.

    The 2 month thing i referred to above, had kpi's built into the sla from all directions and with financial penalties as well, some realistic, a few not.. Not sure of its use, the customer liked it though as it gave them a measure of the efficiency of the contract.

    so my view is the SLA'a can be an instrument of use, that result in total quality management solution, they can be good and bad, but from a client perspective, if the trusting relationship was better, then maybe they are not needed... and then.....

    Having sat on both sides of the fence i can see their merit, butdo we need to re-invent the wheel?

    If suppliers/ service providers were 100% efficient and trustworthy, then lets just work from a PO.

    kind regards, Louise

  8. Thanks Loiuse, It's great to have someone provide a counterpoint to the debate. Personally, I'm a big fan of KPIs rather than SLAs although I use them sparingly (and only for as long as I need them).

    I think SLAs have there place: I wouldn't want to try and run a large TFM, for example, without them. But I do think there are smarter ways to work and plenty of ways to set expectations - I've blogged previously about requirements specification, for example. The important thing is that we will achieve best success (as FMs)as a result of applying the right tool to the job, not by using a multi-tool to do them all.


  9. Just wanted to draw attention to another blog, for those interested in SLAs and KPIs. Check out Tom Robinson's views on their effectiveness in ensuring good customer service. http://www.oneguestatatime.com/blog-2/customer-service-measuring-the-immeasurable