[Discuss] core competency

[dsr at randomstring.org (Dan Ritter)]

On Wed, Jan 23, 2013 at 11:14:13PM -0500, Tom Metro wrote:
> A while back Dan Ritter wrote:
> >Dave Berry wrote:
> >> Does anyone know of companies that offer linux server and website
> >> support? ... Our servers are remotely hosted by a colo company...
> > 
> > There are companies which do that, but let me caution you: do
> > not outsource your core competency. This may be a sign that you
> > need a third person.
> 
> What exactly is a "core competency" for a modern tech company?

It's the things you do that make you money, and the infrastructure that
absolutely must be running in order to support that. 

If you can hire a company to do it for you in exactly the same
way that they do it for other people, it's not a core function.
There are very few companies where accounting is a core
competency, for example... other than accounting firms.

> We went through a period where it was fashionable to recharacterize IT
> from a cost center to being a core competency, but was that based on
> anything rational, or just self-preservation for the IT guys?

Depends on the company. "IT" is usually a cost center, but
Operations -- technology applied to delivering the product or service --
is part of your direct costs of goods sold, and may be the thing
which you are selling.

> Today we outsource big swaths of IT, and it has sort of happened under
> the radar, as you don't seek out and hire an IT consulting firm that
> shows up at the office. Instead, when you need to launch your next
> in-house app., you just rent a cloud server. Eventually, you find that
> most of your IT infrastructure is in the cloud.
> 
> Is this a bad thing? Are you at a competitive disadvantage?

Depends. You are certainly not at a disadvantage compared to other
companies that have done the same thing (which is why buying IBM or
Microsoft or whatever was a conservative move).

If your infrastructure is up when your competitors are down, though,
you have an advantage. That could go either way: when Amazon misplaces
their storage, all their customers are in trouble. When Amazon
gets you up and running faster and initially cheaper than
building your own, that's a win for you.

> But what about all the companies that make software products. Is
> software engineering a core competency of Rovio, maker of Angry Birds? I
> would argue that for most software companies, understanding their
> customers needs and knowing how to design a product that meets them is
> the real core competency.

For every company, understanding customer needs and being able to supply
a product or service to meet them is the primary function. Rovio is not a
great example: they spent six years barely hanging on before they found
a hit. 

> This line of discussion interests me because there is a conventional
> wisdom offered up to startups that it is a mistake for them to outsource
> their engineering.

It depends. Are they relying on engineering as the core of their
business? Will they live or die over whether an engineer is inspired and
dedicated and works overnight fixing a problem or implementing a feature?

If so, then it is a mistake to outsource that function.


> result of such outsourcing. But that may just mean they didn't do it
> well. Or didn't choose a good outsourcing partner (i.e. one that
> understands how to do knowledge transfer and can help build a permanent
> in-house team when the time is right).

I'm pretty sure no "outsourcing partner" is enthused about eliminating
their future revenue stream that way. Recurring revenue is life.

> So is a startup really best served by permanently hiring and building
> its product in-house, which has the very real challenges of finding good
> people, assembling a team that works well together, and then honing a
> development process, all while operating under great resource
> constraints? Most founders underestimate the difficulty of putting
> together a high-performance engineering team. (And a startup usually
> can't afford the mistakes and delays caused by a low-performance team.)
> 
> Wouldn't the startup be better served by outsourcing engineering to an
> experienced manager with a tuned team and process, and instead spending
> its resources on a product manager - or, seeing as that's a role the
> founders will likely handle themselves, more generally product design
> activities?

How are they going to find this manager and team? Once they find them,
why would the startup prefer the uncertainty of their availability and
dedication in the future?

This actually strikes me as a model more likely to work as an experimental
model at an established company rather than a suitable risk for a startup.

-dsr-