Systems development...in house or out sourced?
Systems development...in house or out sourced?
Author
Discussion

thisislife

Original Poster:

344 posts

214 months

Tuesday 18th August 2009
quotequote all
We are currently in the throws of doing a spec for a computer system that we need building. It’s a large system and will be fundamental to our business.

The work will be under constant development so I see this creating a full time job for someone.

My question is....do you think we should have the system made and developed by a systems development company, or should we employ a system developer to work in our offices and build it on our own servers etc.

I like the idea of having a developer in the office so we A) know work is being done and B) can work closely with him/her making sure we are getting exactly what we want.

Your thoughts on this would be greatly received.

Carsie

944 posts

235 months

Tuesday 18th August 2009
quotequote all
It is difficult to offer specific advice as you don't say what busines you are in and why you need a bespoke system

My experience would say that you don't develop a system in house merely customise an application otherwise the pound notes will be flying out of the window faster than a proverbial ferret down a rabbit hole.

There is a business case for developing but is your business that unique? There are a myriad of business systems in the market and your development costs could be managed and controlled.

With regard to Contracting the development,again I would rather see an employee in place as this would give you control not just on financial budget but also scope creep.

Be delighted to advise once you have your Project Outline

Don

28,378 posts

315 months

Tuesday 18th August 2009
quotequote all
Well. There are a few ways you can do it.

1) Develop and build in-house.
2) Buy a product that almost does what you want. There will be one!
3) Go hire some contractors to speed up (1).
4) Or...Go to a systems development house with a spec, get a fixed price quotation.

I run a shop that does (4).

Why you buy (4) instead of doing it in-house is

(1) Fixed price. No risk.
(2) Outsource the software support on an annual basis. People to phone at an office number you always get through to. Usually cheaper than the salary of the support staff you'd need otherwise.

If you do outsource the development and support we'd be happy to help you. UK based. People on the phone who you can understand etc etc.

The benefit of doing it in-house is that the guy who wrote it and supports it is sitting in your offices permanently.

Until he leaves.

Plotloss

67,280 posts

301 months

Tuesday 18th August 2009
quotequote all
thisislife said:
B) can work closely with him/her making sure we are getting exactly what we want.
Having someone in your employ on site is no guarantee that you'll get what you want in my experience.

If you've got the pre-existing capacity within the business to code it yourself then that is probably the most sensible route to go down.

If you have to hire specifically for this work from front to back i.e. no one in the team with technical and business knowledge then you may as well look at a development shop or even an integrator.

sonic_2k_uk

4,008 posts

238 months

Tuesday 18th August 2009
quotequote all
Im a developer and have plenty of experience planning, building and launching the size of system you are talking about from all 3 perspectives (no im not after a job) -

I originally worked for a development agency where we would take a spec from a client like yourself, build and then impliment it.

I moved onto contracting, where i spent alot of time working on in-house projects with existing IT teams to try and meet deadlines, i also spent time developing bespoke systems for client requirements (just like when i was working for the agency).

Now i work as part of an in-house IT team building designing, building and running business critical systems.

The answer to your question is pretty difficult - each avenue has its advantages and disadvantages.

1) Agency - You should be able to get an agreed price for the project providing you can sign off a specification. You will generally get industry professionals with a high level of expertise who's interest is to get the project completed as quickly as possible. You will however not be able to look over their shoulder as they write the code and they will probably not have as much an understanding of the system as an in-house team. Ammendments, additions and deviations to the agreed specification will probably be expensive. You'll me amazed at what certain departments forget they need in the system.

2) Contractors - You should be aiming to get the highest standard of expertise available but it will cost you. The project will generally cost on a time basis, regardless of how long it takes. Developers can range anywhere from £200+ per day, typically around the £300/day mark.

3) In-house - This provides the most flexible cost-effective solution. You can pretty much change whatever you want, whenever you want, although the IT team will seriously not like this!!! Obviously your not paying contractor or even agency rates, only the employee salary which will be less. Typically you will get less skilled staff though and the project will take longer. You will also be responsible for project managing and organising the development, although you may well have some IT manager who takes care of this.

So really it utterly depends on what you need! HTH!

otolith

69,781 posts

235 months

Tuesday 18th August 2009
quotequote all
It also comes down to your organisation's ability to pin down and specify in advance the product that it wants. It's not my preferred way of working, but I have to concede that the open-ended, iterative prototype-and-evolve style of development sometimes works better for smaller systems where there is a full time developer with a close working relationship with the client.

ewenm

28,506 posts

276 months

Tuesday 18th August 2009
quotequote all
Don't forget support. So you've got your shiny new system and after a few months something goes wrong - do you have the skills in-house to fix it? Do you need to call on the original development team (big rates perhaps)? Are you paying a support contract? etc etc.

My gut reaction would be to find an off-the-shelf solution that most closely matches your needs and modify it for your business.

Don

28,378 posts

315 months

Tuesday 18th August 2009
quotequote all
otolith said:
It also comes down to your organisation's ability to pin down and specify in advance the product that it wants. It's not my preferred way of working, but I have to concede that the open-ended, iterative prototype-and-evolve style of development sometimes works better for smaller systems where there is a full time developer with a close working relationship with the client.
RAD interactive development is GREAT for end users as they can change their minds again and again and again. The software evolves with the business.

However it's BAD in terms of getting a solution quickly and it can be hard to draw a line under the development and say "No more! We have spent enough!"

The ideal software development is not to do one at all, to be honest. Just buy a working product that meets the requirement. Bang for your buck is usually best.

Bespoke solutions developed by a pro software house aren't cheap. But they may not cost more than running a project in-house if you end up needing contractors at £300 a day to get the job done in a reasonable time-frame.

It isn't black and white which one is best, either. I will say this, though: if you are in IT yourselves why on earth would you let anyone else build you a bespoke system. If you aren't in IT why on earth would you try and do it yourselves...

Don

28,378 posts

315 months

Tuesday 18th August 2009
quotequote all
ewenm said:
My gut reaction would be to find an off-the-shelf solution that most closely matches your needs and modify it for your business.
This is less easy to do that it would at first appear. The app either needs to be highly configurable (e.g. SAP) or open source (pot luck as to whether there is anything and whether the code is any good and whether or not you can get any expertise in it etc etc).

We've looked at open source stuff and then developed our own before now...because we develop software for a living we, of course, develop our own stuff when we need it - IF we can't just buy a product.

ewenm

28,506 posts

276 months

Tuesday 18th August 2009
quotequote all
Don said:
ewenm said:
My gut reaction would be to find an off-the-shelf solution that most closely matches your needs and modify it for your business.
This is less easy to do that it would at first appear. The app either needs to be highly configurable (e.g. SAP) or open source (pot luck as to whether there is anything and whether the code is any good and whether or not you can get any expertise in it etc etc).

We've looked at open source stuff and then developed our own before now...because we develop software for a living we, of course, develop our own stuff when we need it - IF we can't just buy a product.
Indeed, I work in developing/configuring/enhancing/customising Oracle applications. The most common problem is that the owner of the closest-fit software is unwilling to modify it for your needs, unwilling to allow you to develop it yourself or charges through the roof for doing so.

otolith

69,781 posts

235 months

Tuesday 18th August 2009
quotequote all
Or lies through their teeth about the ease of integration and the support available to do so...

Or fobs you off with something entirely useless *


  • Something dangerous, in fact. Of course you can query our directory system for your Intranet authentication system. Here is a component and some sample code. We'll let you figure out yourself that the threading model it uses will cripple your server if you use it as specified. C**ts - and their LDAP implementation was borked.
An in-house development team is perfectly feasible if you have sufficient need of bespoke software to justify it - doesn't necessarily need to be your core business. The last one I was a member of was audited by a company looking to bid for outsourcing options (so no conflict of interests there) - they told the board they were better off with us (but we'll have your desktop support, ta). The costs of customising ERP solutions aren't always as competitive as the sales staff will tell you, either.

mouseymousey

2,642 posts

268 months

Tuesday 18th August 2009
quotequote all
You may think that you have more control if it's done in-house but that's not necessarily true.

If I.T. isn't your business then you could end up with an an absolute dodgepot developer.

The problem is it can be very easy for an in-house developer to bullst his way along, blaming non-existent problems for slow progress etc, often because the person managing the developer has no idea about software development. You probably won't realise that they are rubbish until they leave, probably just before a big deadline that they've promised is due.

Then the new guy will come in and, assuming they can find the code, they will tell you that it's terrible and huge parts of it need re-writing.

In my experience, if you need to write something in house then it's not just a case of employing a developer. You will also need someone to project manage it and ideally someone to properly test it.

If you outsource it then, if you pick your partner correctly, you will get access to professional project mangers who can help you understand and write your spec and manage the entire project lifecycle. You will get professional developers, testers and support people. Everything will be fully documented and put under source control and version management. This means that when the lead developer leaves the firm that wrote the software you will not notice, you'll get a seamless transfer.

As has been said before, there's no right answer but you need to fully understand all the costs involved in both options. I find that professional outsourced development can look expensive as the costs are all known up front, developing internally may look cheaper on paper but allowing the software to be written badly could go on costing you money long after you think the project should have been completed.

Don

28,378 posts

315 months

Tuesday 18th August 2009
quotequote all
mouseymousey said:
If you outsource it then, if you pick your partner correctly, you will get access to professional project mangers who can help you understand and write your spec and manage the entire project lifecycle. You will get professional developers, testers and support people. Everything will be fully documented and put under source control and version management. This means that when the lead developer leaves the firm that wrote the software you will not notice, you'll get a seamless transfer.
This is a key reason for choosing to outsource to a software development company. yes It's also a reason why the really big companies buy their software development from other really big companies: no one got fired for buying IBM, they used to say until CROESO went tits up.

But there are "lower tier" firms, like mine and others, that offer a more cost effective version of the service.

mouseymousey said:
As has been said before, there's no right answer but you need to fully understand all the costs involved in both options. I find that professional outsourced development can look expensive as the costs are all known up front, developing internally may look cheaper on paper but allowing the software to be written badly could go on costing you money long after you think the project should have been completed.
Agree completely with that.

Don

28,378 posts

315 months

Tuesday 18th August 2009
quotequote all
otolith said:
The costs of customising ERP solutions aren't always as competitive as the sales staff will tell you, either.
rofl Or as quick.

We once put in a system. Designed to be an "interim" solution until an ERP whose name begins with S ends with P and has an A in the middle could be deployed.

Our solution worked and cost less to build and deploy than the annual support for the essaypee system.

Understandably our solution is still in use ten years later... hehe

otolith

69,781 posts

235 months

Tuesday 18th August 2009
quotequote all
mouseymousey said:
... outsourced development can look expensive as the costs are all known up front...
Potentially - a naive company can get still their trousers taken down, and it's not unknown for suppliers to bid low with the full intention of turning their profit on entirely predictable changes and scope creep.

otolith

69,781 posts

235 months

Tuesday 18th August 2009
quotequote all
Don said:
Understandably our solution is still in use ten years later... hehe
hehe

My experience of ERP solutions is that they are sold to the board as a panacea.

thisislife

Original Poster:

344 posts

214 months

Tuesday 18th August 2009
quotequote all
Many thanks for all your comments they were greatly received.

Ill go into a little more detail. We have had a system built by a developer (1 man band). It has cost us in total between 15-20k and has been in development since Feb. It was completed about 1 week ago.

It works....it does everything we require, but i get the feeling it has been bodged. Its not very smooth and the functionality and feel of it is not great.

It has cost a fair few quid and will do us for now but as our company grows I can’t see this system growing with us.

With regards to comments on an "off the shelf system", there are none that do exactly as we require which is why we have gone the bespoke route. Our system is absolutely vital to our business and many of its automated functions cut down manual staff hours massively thus saving us a fortune in wages every month.

One of my main concerns about using an out sourced development team is that because our business is fast evolving so are our needs from the system and our spec is constantly changing. Its more of a "work in progress" situation. I couldn’t honestly sit down with a development firm and draw out an exact spec because in 2-3 months time that spec would have changed and i will be looking at more money and time etc.

If we had someone in house, we could work closely together which I really feel would work. The comments on "project management" and the system working until said employee leaves have got me thinking though....

confused

Fittster

20,120 posts

244 months

Tuesday 18th August 2009
quotequote all
otolith said:
mouseymousey said:
... outsourced development can look expensive as the costs are all known up front...
Potentially - a naive company can get still their trousers taken down, and it's not unknown for suppliers to bid low with the full intention of turning their profit on entirely predictable changes and scope creep.
Hmm, scope creep smile, that's were the margin is. "Oh, you want it to work? That's extra!".

If there is any scale to the system at all the chance are the agreed deliverables will mean completely different things to all parties. The trouble is you won't want to do the level of design necessary to get an properly agreed spec, so tears are likely.

mouseymousey

2,642 posts

268 months

Tuesday 18th August 2009
quotequote all
thisislife said:
Many thanks for all your comments they were greatly received.

Ill go into a little more detail. We have had a system built by a developer (1 man band). It has cost us in total between 15-20k and has been in development since Feb. It was completed about 1 week ago.
This is often how a "one man band" system will feel. It may well work but the chances are that it hasn't been written in the most efficient way, which will become a problem later if it is required to scale up.

Are you able to give us any clue on which industry you're in? I ask from a purely self interested point of view as my company writes software for retailers.



Don

28,378 posts

315 months

Tuesday 18th August 2009
quotequote all
thisislife said:
Many thanks for all your comments they were greatly received.

Ill go into a little more detail. We have had a system built by a developer (1 man band). It has cost us in total between 15-20k and has been in development since Feb. It was completed about 1 week ago.

It works....it does everything we require, but i get the feeling it has been bodged. Its not very smooth and the functionality and feel of it is not great.

It has cost a fair few quid and will do us for now but as our company grows I can’t see this system growing with us.

With regards to comments on an "off the shelf system", there are none that do exactly as we require which is why we have gone the bespoke route. Our system is absolutely vital to our business and many of its automated functions cut down manual staff hours massively thus saving us a fortune in wages every month.

One of my main concerns about using an out sourced development team is that because our business is fast evolving so are our needs from the system and our spec is constantly changing. Its more of a "work in progress" situation. I couldn’t honestly sit down with a development firm and draw out an exact spec because in 2-3 months time that spec would have changed and i will be looking at more money and time etc.

If we had someone in house, we could work closely together which I really feel would work. The comments on "project management" and the system working until said employee leaves have got me thinking though....

confused
We quite often win a product sale in conditions like these. Typically the guy who ran the timesheets system in Access and Excel for reports and a bit of VB so people can fill 'em in leaves. And the whole thing has to be replaced a.s.a.p. At that point people buy a product.

I think you are in a fairly decent position, though.

(1) You have some software that works - it may be bodged - but it works. Which means it doesn't need to be replaced right now. You can take your time deciding! So long as your developer doesn't bugger off...

(2) A working piece of software is a superb specification - your new one MUST do what the old one does!

(3) 15 - 20K is a reasonable investment for a fully working specification of what your solution has to do. I think you've done OK.

Before now we've been hired to "reverse engineer" a working system and to build a replacement. The trigger for doing that is usually something like a platform change: i.e. the old one was built in Access and Microsoft Office and the new one is going to be hosted on a web server in Canary Wharf and use a SQL Server db etc etc