How they know you know

There have been many times over the years when someone has asked me to assess someone's skills. Usually this is by looking at a CV, but it has also involved talking to them in an interview setting.

For any potential employer, the hiring process is painful. They know they need someone, but typically they don't really know who they need. Could be they need to hire someone to join a team, in which case they can ask the team what kind of person they would like. But often they just know that they need someone to do a particular job.

First step, they need to write a job specification. Great. What they really want to write is "We need someone to be able to fix up this stuff" – but they actually have very little idea about what skills may be required for that. So they put criteria like "Must be good with people" and "Must have experience in the integration of systems". But really, they have very little idea.

Let's suppose somehow, they get a terrific advert written up, and the CVs start flying in. That's a really good start. Now they need to go through those CVs and work out which of these people they are going to invest their time in. Getting someone in for an interview will take time. There's the interview time of course, but as well as that, there's arranging a time that suits, and working out a range of questions that not only suit the role, but suit the candidate. They might want to ask about the candidate's time abroad, or why they left that job which sounded so good. All-up, I would guess that a single hour-long interview would take around 3 hours of effort.

So you can't interview everyone, you need to use CVs to filter them out. That's tough. You try to rank them according to a feel for whether one is better than another. But how do you do that? Is this person who has worked for a couple of big name consultancies better than that person who has a degree from a top university? Does it matter that this person has a few gaps in his CV? Lots of questions need to be considered.

Typically as a candidate, you need to be able to demonstrate very quickly that you are better than the crowd. Obviously if you have an outstanding employment history, have written books, that type of thing, then you may get through on reputation alone. But if you have a good reputation, then you're probably not applying for jobs anyway – the employers are probably chasing you directly. For the rest of us, it's much harder.

Certification can help I think. If someone gets two identical (-ish) CVs across their desk but can only get one in for an interview, does it help that one has achieved some level of certification? I think so. Does it help that one is a member of a professional society? I think so. It doesn't mean that the candidate is necessarily any better, but it ought to increase the odds of it. The person without the certification might not be good enough to get certified. Or they might not qualify for membership of a professional society. So I think employers will always lean towards the person who has these things.

When you get a certification from Microsoft, they send you a certificate in a folder. The folder says "How they know you know". And I think this is the key. Whilst certifications don't guarantee you know anything, I think they do go a long way. And after all, even if it's more like "How they can justify hiring you rather than the other guy", then surely that makes the certifications worthwhile.


I'm on the MS Developer Podcast

I'm listening to myself right now on the The Microsoft Developer Show from The Podcast Network, being interviewed about certification. The link is

I do talk too fast – that's a pain. I really need to work on that, especially if it's being recorded for a podcast. I don't talk so fast on the ones that I record myself (but they're not technical, and I'm just talking to myself, not to someone like Nick).

The show notes scare me a little. I see a mention of Dave Lemphers at 4:01. I like Dave a lot – great guy. When I saw his name there though, I thought "Oh, I don't remember mentioning Dave…", and wondered what it was about, considering some of his views on the MVP program.

Having listened to it, it's not too bad, except that I talk too long at the end (what – Rob talking too much? Can't be true…).

I'm happy to promote certification, and also to encourage people (both candidates and employers) to think higher of the certifications. After all, we really do need to know how people know.

SQL Talk Drinking game

Personally, I don't drink. Not at all. Just one of those things. But, at TechEd this year, I thought of a drinking game while listening to people talk about data.

In the UK, people pronounce data as 'dayta'. In the US, it's 'datta'. And as far as I can tell, people from these places don't tend to vary which one they use. But in Australia, the typical pronunctiation is 'darta'. Yeah, I don't get it either. But that's just the way it is.

Now, because of the international nature of the community, many Australian SQL presenters have started to say 'dayta' instead. No-one says 'datta', but people switch between the other two even within the same sentence – (putting 'dayta' into a 'dartabase')

So in my drinking game, you split into pairs. One person is assigned 'dayta', and one is assigned 'darta'. And of course, each time the presenter says it each way, the appropriate person takes a drink (very small, or else they get very drunk very quickly).

SQL Code Camp

Well, here I am at Wagga Wagga. It's Sunday morning, and Itzik Ben-Gan is speaking. He's just shown us slides of the SQLHike trip, which looks like a really cool idea.

Being here in the middle of nowhere talking about SQL with people is really fun. My talk (on the OVER() clause) went okay yesterday. I was last up and had to keep it short, but I got good feedback from people. People liked it when I jokingly told Itzik to just put his hand up if he had any questions. I must check out Itzik's talk on row_number() some time – I'm sure I'd learn plenty, and be able to make my talk even better.