So You Want to Set-Up a Database?
By Margot Lunnon
Your existing database is old and cranky and needs replacing or you're about to take the plunge and set up a database for the first time. So how do you go about it? Can you buy a package off-the-shelf? Should you go in for software development? Can a consultant set up something for you at a reasonable cost? Lasa's Information Systems Team and freelance database consultant Margot Lunnon look at three possibilities: buying off-the-shelf, full scale development, or a cheap and cheerful solution from a consultant.
It should be relatively speedy process. Once you've located a suitable product you should be able to get it installed and begin the process of implementation without too much delay. A well established product will have been used by previous customers and incorporate a wealth of their experience. It may incorporate new ideas and ways of working, and lead to you improving your processes and practice.
You should have the benefit of a telephone help-line to deal with any technical difficulties and you may even be able to get some help on the best way to make use of the product. The company you've bought from should have accumulated a lot of knowledge about the way their customers use the system and may be able to offer guidance in how to achieve what you want. It should be a tried and tested product. You won't be the first person using it and won't be testing for bugs. And it should be cheaper.
There are disadvantages: You may have to accept an element of compromise when selecting your software. It may not do everything you want or do it exactly as you would like. This is a bit dependent on how unique your requirements are.
You probably won't be able to find an off-the-shelf system, if what you do is very unusual, but there should be a choice of packaged solutions for a common activity. It can be quite hard to find the product you want. You can try publications like the Software Yearbook in the local library but it may not include specialised voluntary sector systems.
Try asking around other organisations in your field to find out what they use. It can take quite an effort to find a suitable product and to satisfy yourself that it is suitable. Some off-the-shelf systems are from very small companies.
It is common for software companies to develop a system for one client and then market it on as an off-the-shelf package. You don't want to be the second user, and you don't want to buy from a company that is about to go bust. Check the viability of the vendor very carefully. You want a company that is growing with a big enough market to survive. Check their ability to keep their product up-to-date - how long has their Windows XP version been available?
An off-the-shelf system should be the priority for most organisations. Only once you're sure it isn't right for you, perhaps because your requirements are very particular, should you look at the more demanding option of software development.
The Full Monty: consultancy
In this context the Full Monty means having a full consultancy, having a database built for you by a specialist.
This might cost only a couple of thousand, but it might cost £10,000 or more, depending on the work to be computerised and the degree of automation and error-checking to be incorporated.
If you are going down this route you should expect it to take time…
Before the work even starts, there will be time to prepare the brief, time to meet candidate consultants, time to refine the brief. Once the work has started there is more time, because the consultant selected will want more meetings to enable him or her to get very clear indeed about the work to be computerised, how it is done now, how it might be done better.
All these meetings are gruelling, they require a an unusual degree of thoroughness, and they usually involve several people. Once a usable version of the database is installed the consultant will expect the client to test it exhaustively - and so it all goes on.
Full consultancy is not invariably the safest option, although it ought to be. It ought to be, because the specification for the work should spell out exactly how the system is to behave once it is in and working, and the clients ought to have taken up references for the consultant with some care before the work started. Anyone who accepts a superficial spec or doesn't follow up references almost deserves what they get, and what they get is commonly a waste of money.
However, an organisation may be at a stage where it is not equal to the full spec process, and this may be for good as well as bad reasons. The work may be very new (if this is the case off-the-shelf software may be the best choice so long as it is cheap). Or, senior workers may be so engrossed in other work that they will not give the time to the specification process that it really deserves. (They may become willing later once they have used a rough-and-ready database!) Or the organisation may be unable to find the funds for the Full Monty and unwilling to wait until it can.
Quick and Dirty
By Quick and Dirty what I mean here is a database built by a specialist but on a more rough and ready basis than that described above. It may be the right solution when a worker has been on a beginner's database course and has been warned by the trainer that building databases is not child's play, especially when they involve a multiplicity of related tables.
The Quick and Dirty jobs that I have done have cost in the order of £800, at least at the outset (Organisations have sometimes come back to ask for more as their working situation has developed.). This has paid for about three full days' work, perhaps spread over more visits. I tend to work largely or entirely at the client's workplace so that I can check with the client continuously as questions arise. I also work as hard and fast as I can in these circumstances.
It has proved, however, that we must not dispense with a spec, even though it obviously has to be an outline document. I have to understand the work being computerised, I have to know what the client is counting on me to deliver. When I have gone back for a follow-up the client will often have a list of tasks, but even these will have to be prioritised in case we run out of time. Because the worker can be beside me as I work, he or she may learn from watching; and indeed in one case I didn't so much build the database as provide the worker with a range of techniques which she could use to build it.
I do not want to give the impression that one gets from a Quick and Dirty all that one might get from a fully spec'ed database. Something is bound to get left out. The sorts of things that may get left out are these:
- there may not be error-trapping to complain if, for example, the keyer enters a date that makes no sense.
- automation probably won't get built -where, in life, one transaction does not happen without another happening too, in the database both may have to be keyed separately
- documentation will be sketchy and so will training.
Database development remains a key problem for voluntary organisations. Anyone thinking of using IT to manage information will be faced with the task of setting up a database.
Many people underestimate the size of such a project, and get a considerable shock when they realise the cost and skills needed to do it successfully. We continue to get helpline calls from agencies which have got funding for a new information project based on the assumption that they can buy a suitable package for a few hundred pounds.
The trouble is that the complicated information needs of voluntary organisations are not so easily satisfied. We need databases tailored to our particular and quite demanding needs, and such systems are not so readily available.
The voluntary sector market is small and poor and so it's hard to recover the high cost of developing top quality systems. More specialist products are becoming available, most funded by public money, but the development process is hard and it is difficult to sustain the income needed to refine and upgrade these systems.
There's an interesting comparison with website development. Here, organisations can use mass market web software: there's nothing very different about a voluntary sector website. The result is that web development is relatively straightforward, and organisations can develop sophisticated sites at reasonable cost. In contrast the complexity of information needs and the lack of specialist software mean many organisations struggle when faced with the need to set up database systems. Database development is very much the last frontier.
For more on getting a database see other articles in the database section of the knowledgebase and the Techsoup article Should Nonprofit Agencies Build or Buy Databases.
Published: 15th March 2002 Reviewed: 9th April 2006
Copyright © 2002 Margot Lunnon
All rights reserved