Buying IT Services: the series
PicNet has sold IT services to Australian organisations since 2001. Over that time we’ve provided managed support, cloud hosting, custom software, integrations and, more recently, production AI systems. Twenty-five years on the supplier side of the table shows you where buying goes well and where it goes badly. This series is our attempt to write that down for the people doing the buying.
Who it’s for
The series is for IT managers, CIOs and the finance or operations executives who end up choosing a provider, signing the contract and then living with it for years. Some of you run a sizeable internal IT team and want a second opinion on a sourcing decision. Others are the person in a 40-person business who looks after IT on top of a day job. Both groups have to make the same decisions, usually with less information than the supplier has.
What it covers
The posts deal with the practical questions that come up when you pay someone else to do IT work. We’ll look at how to read a managed services proposal and work out what’s actually included. We’ll cover what fixed-price and time-and-materials arrangements protect you from and where they leave you exposed, what a sensible service level looks like for an organisation your size, and how to find out whether the engineers in the pitch meeting will be the ones doing the work.
AI is now part of many IT budgets, and buying it comes with its own traps. A lot of AI work looks convincing in a demo and falls apart once it meets real data and the security requirements of a working organisation. We build production AI systems, so parts of the series cover how to scope and assess that kind of project too.
How we’re approaching it
We’ll write the way we’d talk to a client across the table. Where we can give real costs, we will, including the ones that tend to appear after the contract is signed, like transition work and the hours your own staff spend managing the supplier. We’ll also be plain about limitations, ours included. A small senior team suits some organisations and doesn’t suit others, and it’s better you know that before you engage anyone. When the honest answer is “it depends”, we’ll explain what it depends on.
Every manager at PicNet is an engineer, and that shows in what we pay attention to. The posts lean towards how things behave in production: what happens when a server falls over at 2am, or what an integration costs to keep running after go-live. We think those details tell you more about a supplier than the proposal does.
We are a supplier, so keep that in mind as you read. We’ve tried to make the series useful whether or not you ever talk to us, and when a point happens to favour the way we work, we’ll say so.
PicNet builds production AI systems for Australian organisations. If you’re working out where to start, talk to us about what a first project could look like.