Tag Archives: G-cloud

An Open Letter to Software Suppliers – 13 Ways to Help the Public Sector to the Cloud

Dear Software Supplier,
The public sector is hemorrhaging money. Money
We all use the same systems, but instead of joining up and sharing we are each deploying identical solutions in isolation.
This is costing us a fortune at a time when (by the way) money’s too tight to mention.
A first step to joining things up is to get all this software out of our data centres and in to the cloud.
We’re trying to get to the cloud – we really are – but you’re not making it easy for us. Here are 13 steps you can take to facilitate our journey to the cloud whilst simultaneously making yourselves some money.
1. Be Cloud Ready
This may seem odd coming from a Head of ICT – but we don’t want to have our own IT departments.
We don’t want any tin on-premise. The public sector should not be in the IT business. An analogy – we consume electricity but we don’t run our own power station. We consume data – but we don’t want to run our own data centre – we want Software as a Service (SaaS).
At my local authority we’ve had a ‘cloud first’ policy since 2007. Every time we buy an application or renew a contract we try to get the vendor to host it for us. This is harder to do than you might think.
More often than not the question “Will you host it for us?” is greeted by the exchange of bemused/panicky looks between the vendor’s sales folk.
“Um, we could install some servers in your data centre and then look after it remotely?” Noooooooooo!
Vendors need to understand that, within a year or two, if you can’t deliver your application as SaaS then we are not buying it.
Vendors should be cloud ready – wshouldn’t have to start from scratch every time. I want you to spin up my instance of your application at the touch of a button. So you’ll need to…
servers
2. Invest Upfront in the Platform
You need to have your infrastructure built and ready to go before you enter pre-sales.
Regardless of whether the servers are in your own data centre or Rackspace’s – they should be humming away before we sign the contract.
The tin is not the customer’s problem.
3. Avoid Short-termism (understand our business case)
There are many reasons why cloud/SaaS is attractive to the public sector – perhaps the foremost being that SaaS should be cheaper than on-premise. I say ‘should’ because it usually isn’t.
We tried to move 2 of our largest, mission critical, systems to the cloud last year. These systems are sold by 2 of the largest software houses in the world. In each case the total cost of ownership over 5 years was TWICE as high under a cloud/SaaS model when compared with traditional on-premise hosting. So we bought the tin, put it in our data centre and it’ll now be at least 5 years before we have an opportunity to look again at the cloud for these systems. Frustrating.
The reason that these two vendors were unable to make their SaaS offering compete with on-premise is that they did not take the long view.

nepho

As with most services, the bulk of the cost of any IT department is in salaries. In theory IT departments should be able to reduce headcount as a direct result of moving systems in to the cloud. This, of course, is why many IT teams have nephophobia (fear of clouds). IT departments think clouds are evil. 
The problem is, though, we can’t reduce headcount as a result of just a couple of systems going SaaS. If we put some of our databases in the cloud we’re still going to need DBAs because we still have lots of databases. This is a major challenge to the SaaS business case.
What the vendors need to do is reduce profit margins for these early SaaS forays. By taking a hit now they will be paving the way for organisations to put more of their systems in the vendor’s cloud. You need to be at least matching, if not beating, the on-premise cost – otherwise, why would we do it?
Short-term pain for long-term gain.
4. Be Secure
It seems that there are a surprising number of software companies who don’t know much about IT security, PSN controls or the Data Protection Act (DPA).
You need to be PSN experts. You must understand this stuff inside out.
Golden Gate BridgeThe UK DPA states that our data must be stored within Europe. IF YOUR SERVERS ARE IN CALIFORNIA THEN WE CAN’T DO BUSINESS. 
We spend an inordinate amount of time verifying that the vendor’s cloud is secure – endless surveys and toing and froing – sometimes we even have to visit the vendor’s data centre. On one occasion we found that a very large supplier intended to host our sensitive, precious, data in a crumbling Victorian warehouse by a river. No thanks.
You need to get together with other vendors and with Cabinet Office and come up with an accreditation scheme. Some kind of SaaS ‘kite mark’. All we want to have to do, as the customer, is ask to see a copy of your certificate and be comforted that you know what you’re doing.
5. Build in Disaster Recover as Standard
One of the big attractions of the cloud is that it is immune to localised emergencies. Yet it is surprisingly common to find SaaS offerings that don’t include back-up to a second location as standard.
We shouldn’t have to ask whether you back-up our data to a geographically distant location – and we certainly don’t want it included as an optional (chargeable) extra. It should go without saying that you’ve thought about disaster recovery and this would be a condition of getting your SaaS ‘kite mark’.
 

6. Make Updates/Patches/Releases Opaque to the Customer

When Google adds a new feature to Gmail, or whatever, they do it without any fuss. We generally don’t know that an upgrade has happened until after the fact.

The same should be true of SaaS releases. This work needs to happen behind the scenes without the need for any downtime.
Similarly, improvements and innovations that you’ve developed for one customer should be quickly shared with all your customers (at no extra cost).

gcloud

7. Get on a Framework
Make your product easy to procure by getting yourselves on a framework such as G-Cloud. Public Sector procurement rules are hugely restrictive and prescriptive – so a product that is easy to procure is very attractive to us.
8. Think About Our Capex vs Opex Problem
Many public sector ICT projects are funded using capital monies – often from prudential borrowing.
We can only spend capital money if we can demonstrate that the investment leads directly to the creation of an asset (tangible or intangible). This is fine for a traditional on-premise delivery model because it’s easy to demonstrate that a tangible asset has been created.
But we can’t use capital money to fund the creation of a SaaS solution because there is no asset being created. At the end of the contract there is nothing that the public sector organisation ‘owns’. Sometimes this isn’t a problem as some organisations will find it easier to lay their hands on revenue (opex) monies, but sometimes it can be a deal breaker.
You can help us here by thinking of ways in which the Saas deal can lead to the creation of an asset that the customer owns. Could you, perhaps, nominate a piece of your cloud infrastructure as belonging to us, the customer, and write this in to the contract? In reality, should the relationship come to an end, we probably wouldn’t want to go to the trouble of availing ourselves of this clause – we don’t want the tin – but the clause’s existence may be enough to convince our accountants that this SaaS project is a genuine candidate for capital investment.
9. Make Short Contracts More Attractive
The days of 5 and 7 year contacts are over. New disruptive technologies mean that we need to be able to react faster than ever before. Look at the way the iPhone has changed the

bohrlandscape – we have contracts that are still running that were signed before anyone had ever heard of an ‘app’.

Predicting technology futures is harder than ever – we don’t know what we’ll need in 3 years time so we don’t want to sign 3 year contracts.

I appreciate that this is a commercial challenge – you’re losing a guaranteed revenue stream – but you need to find a way to make short contracts attractive to us.
10. Encourage Multi-Organisation Contracts
This is our responsibility as much as yours – but it’s important to construct contracts in such a way that other public sector organisations can get on board at a later date without going around the procurement rigmarole again.
When selling your wares you should encourage your customers to contact their partners/neighbours to see if they would like to be cited in the contract as part of a procurement consortium – they’ll thank you for it.
11. Be Device Independent
devices
Hopefully this goes without saying these days, but mobile is king, and whatever your SaaS product is for, it should work just as well on an iPad as it does on a laptop.
12. Be Open
Let us at our data! We’ve got big plans for big data so you need to build your cloud offering with openness in mind. We want APIs included as standard that allow us to easily extract data to work on elsewhere.
13. Allow for Online Ratings from Customers
I’ve never come across a SLA that, when push comes to shove, was fit for purpose. Most SLAs are carefully worded such that it’s very hard for them to be breached and that the criteria for triggering penalties is rarely met.

onlinerating

The purpose of contracts and SLAs is to drive a certain kind of behaviour – ie to increase quality, responsiveness and uptime. But a SLA is a crude tool for this. Much better is a system of open online reviews and ratings by your customers.
This is scary, I know – but your product will definitely get better when everyone is able to talk about it openly.

I appreciate that there’s a lot of pain involved in getting yourselves cloud-ready – but if you don’t do it soon some disruptive new SaaS player will come along and take your business. By the way, if you are that disruptive new player – please get in touch – we’ve got some money for you.

Yours sincerely,
The Public Sector