The latest piece of craft from Kat Hall on how a “GDS Monopoly leaves UK.gov at risk of IT cock-ups” was interesting, to say the least. I’m sure Kat Hall is under pressure to write articles, I’ve seen the Register help create some very fine tech journalists (see @mappingbabel) and I have no doubt Kat will follow the same path. However, this one instance was not a finest hour.
I’ll leave it at that though because my real interest lies with the report and not a debate over "what is journalism". Before promoting a report, I tend to ask some questions - why was it written, why now, who wrote it, what was it based upon and how will it help? I do this because a lot of stuff in the ether is just PR / lobbying junk dressed up as helpful advice.
At this moment in time (due to the election), Civil Servants are governed by the Purdah convention which limits their ability to respond. What this means is that any old lobbying firm can publish any old tat knowing they’re unlikely to get a response. Launching an attack on a department at this time is about as cowardly as you can get. These people are public servants, they work hard for us and a bunch of paid lobbyists or consultants taking swipes is not appropriate.
The report “GOVERNMENT DIGITAL SERVICE 2015” was written by BDO. They’re a big consultancy working in the public and commercial sector, with a glossy web site and lots of pictures of smiling, happy and clapping people. They talk a lot about innovation models, exceptional client service and “value chain tax planning”.
The report starts with GDS has been an effective catalyst for transformation (basically, let's be nice to you and pretend we're friends before we bring the punches out) and then goes on to proclaim major risks which need to be sorted! I’m already starting to get that icky feeling that “major risks which need to be sorted” is code for “pay us lots of money”.
Ok, the three major risks are highlighted as accountability, commercial and efficiency. We will go through each in turn.
THE ACCOUNTABILITY RISK:
“GDS’s hands-on approach to advising programmes reduces its independence as a controls authority”.
A bit of background here. Many years ago, before writing the Better for Less paper, I visited a number of departments. All these departments suffered from excessive outsourcing i.e. they had outsourced so much of their engineering capability they were unable to effectively negotiate with vendors as the department was often little more than project managers. In the Better for Less paper we talked about the need for intelligent customers, that the current environment had to be rebalanced, and that we had to develop skills again in Government. Now, this excessive form of outsourcing wasn’t a political dogma but a management dogma. It’s why we used to be paying through the nose for stuff which wasn’t often fit for purpose. With a bit more internal skill, I’ve seen £1.7M contracts tumble to £96,000. Yes, 95% savings are not unheard of.
However, it's not just GDS. There’s many Departments, the Tech Leaders Network and systems like G-Cloud which have made a difference. A very important factor in this was OCTO (Spend Control) and their introduction of a policy of challenging spending.
The report says “Accountability is the key to risk management and accountability must always be with the department that holds the budget and is mandated with the service” and that has always been the case. The Departments are accountable and they hold the budget. However, CHALLENGE is an essential part of effective management and that requires the skills necessary to challenge.
To explain why this is important, I'll give you an example from a Dept / Vendor negotiation which in essence was little more than :-
Dept. “What options do we have for building this system?”
Vendor “Rubbish, Rubbish, Us”
Dept. “Oh, we better have you then. How much?”
Vendor “£180 million”
Dept “Could you do it for £170 million?”
Vendor “Ok”
It wasn’t quite like that as the vendor had to write some truly awful specification documents and options analysis which it charged an eye watering price for under a fixed preferred supplier agreement. There was a semblance of a process but no effective challenge. You couldn’t blame the department either, the past mantra had been outsource all and they didn't have the skills to know what was reasonable. I’ve seen exactly the same problem repeated in the commercial world numerous times - departments operating in isolation, alone, without the skills required. They are easing pickings.
GDS and Spend Control changed that by forcing some challenge in the process. Of course, if you’re used to chowing down on Government as an easy lunch then those changes probably haven’t been very welcome. Whilst some Departments were bound not to like being asked hard questions - “but, it’s our budget” - others responded by skilling up with necessary capabilities.
You can’t separate a control authority (the point of challenge) from the skills needed to challenge unless your goal is to pay oodles of cash to outside vendors for poor delivery. I can see the benefit for a consultancy delivering services but not to a Government serving the public interest.
THE COMMERCIAL RISK:
“GDS’s preference for input based commercial arrangements rather than a more traditional outcomes-based commercial approach”
First, as someone who created outcome based models for development a decade ago then I can clearly state this is not traditional unless the outcome is delivery to a specification document. This is an important distinction to understand.
One of the key focus of GDS has been on user need i.e. identifying the volume of transaction Government has, identifying the user needs of those transactions and building to meet the user need. This is a huge departure from the past model where the user need was often buried in a large specification document and the the goal was delivery to the specification whether it met user needs or not. So, you first need to ask which outcome are you focused on - user need or delivery to a specification?
When you are focused on user need, you soon realise you’ll need many components to build that user need. Some of the components will be novel and some will be industrialised (i.e. commodity like). The methods and techniques you will use will vary. I could give examples from the Home Office and others but I’ll use an example map from HS2 (high speed rail) to highlight this point.
Example map
Example map
The user need is at the top. There are many components. The way you treat them will be different according to how evolved those components are. This sort of mapping technique is becoming more popular because it focuses on efficient provision of user needs. Doing this involves multiple different types of inputs from products to utility services to even custom built components and applying appropriate methods.
Now, in the traditional approach which is building to a specification then there is usually very little description of the user need (or where it exists it’s buried in the document) and almost certainly no map. This delivery mechanism normally involves a very structured method to ensure delivery against the specification i.e. the focus is not “did we deliver what the user needed” but “did we deliver what was in the specification / contract”. Consultants love this approach and for good reasons which I'll explain.
Take a look at the map from HS2 again. Some of the components are in the uncharted space (meaning unknown, novel, constantly changing) whilst others are more industrialised (well defined, well understood, common). Whilst the industrialised components can be specified in detail, no customer can ever specify that which is novel and unknown. Hence, we tend to use methods like six sigma, detailed specifications, utility services and outsourcing for the industrialised components of the project but at the same time we use agile, in-house development for the novel & unknown.
Oh, and btw the maps I use are a communication tool between groups. With the sort of engineers you have a GDS and other Depts then this sort of thinking is often just second nature. You use commodity components / utility services and products where appropriate. You build only what you need and you use the right approaches to do so.
Oh, and btw the maps I use are a communication tool between groups. With the sort of engineers you have a GDS and other Depts then this sort of thinking is often just second nature. You use commodity components / utility services and products where appropriate. You build only what you need and you use the right approaches to do so.
The beauty of forcing a specification document on everything is you force the customer into trying to treat all the components as the same, as though everything is industrialised. You are literally asking the customer to specify the unknown and then you crucify them later on through change control costs. The vendor can always point the finger and blame the customer for “not knowing what they wanted” but then the reality is they couldn’t know. The massive cost overruns through change control are not the fault of change but instead the structured process and the use of specifications where not appropriate.
Hence you have to be really careful here. If someone is asking you to sign up to an outcome based traditional model which in fact means delivery against a defined specification document for the entirety of a large complex system using a very structured process THEN you’ll almost always end up with massive cost overruns and happy vendors / consultants.
I have to be clear, IMHO this is scam and has been known about for a long time.
I have to be clear, IMHO this is scam and has been known about for a long time.
So which way does the report focus? The reports talks about documentation, highlighting the example of MPA and promotes pushing control to CCS (Crown Commercial Services). Hence we can be pretty confident that this will break down into specification documents. It argues “While GDS focuses on embedding quality staff within programmes, MPA pursues more formalised and documented processes” and then it promotes the view of MPA as the solution.
This argument is not only wrong, it is mischievous at best. GDS focuses on user needs and using high quality staff to build complex projects. It does a pretty good job of this and its output is functioning systems. MPA focuses on ensuring the robustness & soundness of projects that are undertaken. It does a pretty good job of this and its output is formal documents. You can’t say “they write documents, we like specification documents and therefore you should use those sorts of documents” as the context is completely different. Some parts of a large complex projects can and should be specified because they are known. Others parts are going to have to be explored. Some parts will need an outcome based approach. You're going to need good "quality" engineers to know and do this along with specialists in procurement to support.
The report then adds another twist - “As a matter of urgency, in order to manage commercial risk, all commercial activities within GDS should be formally passed over to the newly transformed Crown Commercial Service (CCS)”. Let us be clear on what this means. In all probability, we're going to end up forcing specification documents (an almost inevitable consequence of trying to get 'good value' from a contract) even where not appropriate and hand it over to procurement specialists who are unlikely to have the necessary engineering skills to challenge what the vendors say. This is exactly what went wrong with the past.
IMHO, a more honest recommendation would be “As a matter of urgency, Gov should start handing over large wads of cash to us, preferably in a truck”.
For reference, if you want to know how to deal with a complex system then once you have a map, I find the following a useful guide. Please note, that for both methods and procurement techniques then multiple methods are needed in a large complex system. This is also another reason why you map in order to break complex systems into components to treat them effectively. I cannot reiterate how important it is to have purchasing specialists supporting the engineering function. You don't want to lose those skills necessary to challenge. NB the diagram is not a replacement for thought, it's just a guide.
Methods & Purchasing.
THE EFFICIENCY RISK:
“With a monopoly position and a client-base compelled to turn to GDS for advice, there is a risk that they could become an inefficient organisation”
Should we roll the clock back and see what it was like before GDS and talk about inefficient organisation? I think Sally Howes, the NAO's executive leader, sums it up very politely with the statement “the government, Parliament and my own organisation, the NAO, were very aware of how the old fashioned world of long, complex IT projects limited value for money”.
To put it bluntly in my language, we were being shafted. We're nowhere near the end of the journey with GDS and the report completely ignores how Departments are adapting and growing capabilities. There's not much I can find to like in the report, some bits did make me howl though.
I loved the use of “proven methods” in the paper followed by “excellent opportunity for CCS to show that it can meet the needs of a dynamic buying organisation”. So basically, we believe in evidence and because of that statement we recommend you experiment with something unproven and smells a lot like the past? Magic.
However it is only surpassed by “This paper has no evidence to suggest that GDS is too big or too expensive to achieve its aims” which followed a rant on “ Is this meeting the needs of the government departments or is this excessive? Are they the right staff? Are they being paid enough? Do they have the appropriate skills?”
That’s consultant gold right there. I’m going to create a whole bunch of doubts about a problem I’ve no evidence exists in order to flog you a solution you probably don’t need. Here, have my wallet - I’m sold!
The paper then goes on to talk about “To ensure market-driven efficiency of the remaining advisory function, this paper recommends that the advisory function form a joint venture with the private sector, allowing it to grow fast and compete for work alongside other suppliers”. Hang on, we have G-Cloud, we have GDS, we have growing Departmental skills and we should hand advisory to the private sector because it previously provided “limited value for money”?
I’m guessing they are after more than one truck load of cash. I’m pretty sure this isn’t the “high level vision of the future” that the Government is after.
Now don't take this post to mean that GDS is perfect, far from it. There’s plenty of good discussion to be had about how to make things better and about how departments can provide services to other departments. There has been some misinterpretation (e.g. the Towers of SIAM) and there has been some oversteering (e.g. a tyranny of agile) but that’s normal in such a complex change. The achievements already have been pretty remarkable but no-one should be under any illusion that it can’t be better. It can.
However reasonable discussion or debate doesn't involve a consultancy publishing a report flogging a bunch of dubious and outdated methods - let’s take skill away from challenge, lets hand over advisory to private sector, let’s focus on specification documents - as solutions to risks which aren't even quantified. There's nothing to debate, it's just mudslinging. I'm guessing that's why they published it at a time when no-one could respond.
But what about the motivations of the authors? I see one is a head of government consultancy practice and so is the other. I’m guessing they’re hoping to be on the advisory board and paid handsomely for such pearls of wisdom.
I note that Andy Mahon has “wide experience in public sector procurement” gained from his 28 years at BDO, Grant Thornton, KPMG and Capita covering initial business case to PFI. I’m not convinced that someone with so much experience of flogging to Government and working for a consultancy flogging to Government can ever be considered impartial when it comes to advising Government on how not to be flogged.
Now Jack Perschke is a different matter. A long background in different areas plus also he worked for ICT reform group and was a Programme Delivery Director for Student Loans Company Transformation Programme. Well, this report is a bit odd - given his background.
From the minutes of the Student Loans Company (though Jack had just left), the board even took time to praise GDS noting “the engagement with Government Digital Services (GDS) had been very helpful” and “GDS had improved the understanding of the work required, particularly around the build/buy options”. Further minutes talk about ongoing discussion, challenge and support e.g. from “responding to the conditions set by the Government Digital Dervice (GDS), including the benchmark for Programme costs” to the Board noting that "GDS were a key partner in the Programme“.
Surely this is how things should work? I’m surprised Jack Peschke didn’t see that. I can't see how you'd conclude this was a bad thing.
Well, if there is some good to come from the document, some silver lining then IMHO this document provides further indirect evidence of why Government should develop its own capability, skills and situational awareness throughout GDS and the departments. These sorts of reports and outside consultancy engagements rarely bring anything of value other than for the companies writing them.
I think my “major risks which need to be sorted” is code for “pay us lots of money” is about spot on.
I'll come back to this next week as I want to see what else crawls out of the woodwork here. I don't like civil servants being attacked especially by self interested outside consultants at a time when civil servants can't respond.