In this week’s article, I’ve captured 11 steps for SMEs who need to replace some or all of their key business systems. This is what the corporates spend millions of dollars on, and they get it wrong more often than not. But if you follow these steps – carefully and systematically – you might just manage to be one of the success stories and position your business for rapid growth.
Find out what systems the leaders in your industry use…
- Talk to everyone who is willing to talk to you!
- Google for systems, applications and industry bodies who may be able to help you.
- Visit other similar sites and companies who might be doing this better than you.
Identify the key processes (eg. invoicing, sending quotes, sending newsletters – whatever it is) that the new system MUST be able to do from end-to-end, so that you can throw the old system(s) away. Then, divide your people who do these processes into two groups:
1. Those who can think beyond what they do now, and might be able to invent a new process.
2. Those who aren’t capable of changing what they do without being explicitly trained.
Get daily business news.
The latest stories, funding information, and expert advice. Free to sign up.
The first group should form the basis of your project team. The second group should be tasked with documenting exactly what it is that they do now. Take a few days to workshop and document your processes as best you can, and in as much detail as you can.
Develop a shortlist of possible applications that deliver as many of the above processes as possible. Once you have one or – at most – two standout applications:
- Thoroughly research available product information to ensure that the product aligns with your goals, and solves your big-picture problems.
- Identify the particular processes that the product has a good reputation for delivering. Also highlight any other important (to you) processes they claim to include, but might not be so good at. These are the ones you must test thoroughly to see if they will work for you.
- Research the vendor, ie. will they be in business in five years time?
- Find examples of relevant implementations in your industry.
Get a trial version of your preferred application installed into a test environment. Pay for this if you have to, but don’t skip the step.
Test your core business processes using sample data, and document how you are going to do ‘business as usual’ using the new system. Minimise changes to your current business processes wherever possible, and save improvements for later. I realise you’re investing this money to make improvements, but if you cut over to your new system without ensuring that people can do their essential job, then you’ll be in major trouble.
Plan your data migration. Take an extract, and do a test import. Be prepared to throw this out and start again. Whatever issues are thrown up by this – and there will be many – fix as many as you can in your current system and work out a strategy for overcoming each of the other problems.
Plan your training. Hopefully your new system has good, free online training. Plan to run workshops and training sessions. Get key people involved in the testing then use them to train. If people know how to use the system from the outset, you’ll massively improve your chances of a successful outcome. And, in preparing your training, you will highlight potential areas where the system simply doesn’t work as you expected – which gives you a chance to fix things before you go live.
Create an implementation plan. Simply put – write down what needs to be done, in what order and by whom. Agree what a successful “go live” looks like. And how you will “back out” if it fails.
Plan for a “bedding in” period. Capture everyone’s ideas for improvements, but first focus on fixing things that are truly broken. Save the improvements for later, unless they deliver “no brainer” benefit at low risk.
Don’t forget to decommission your old system(s), or make them “read only” to stop people from accidentally forgetting to use the new system.
When the new system is stable, start planning and implementing all those improvements you bought this system for in the first place.
A WARNING FOR SMEs THAT AREN’T IT SAVVY
If – like many – you have an “IT guy” or clever office manager looking after you, you should ONLY put them in change of your project if:
1. They have demonstrably followed this entire process before; and
2. They can explain to you how they did it and what they learned.
My experience is that most “IT guys” and clever office managers have NEVER done this, and you are in no position to have them learn it at your expense. It’s like putting a dental nurse in charge of jaw surgery, simply because she has ‘dental’ in her job description name or thinks it will look good on her CV.
You need to assume they have about the same knowledge and understanding as you do (ie. not much), and get someone experienced in IT business systems to help you. Yes, it’s expensive – but the cost of failure is much, much higher. Try predicting the cost of losing unhappy staff and customers as a basis for your budget for getting help. Also factor in the cost of your best, most competent staff being focused on systems replacement instead of their real job for five to 10 times longer than necessary, because they don’t know what they don’t know, and can’t explain why it simply isn’t going to plan.
And – my final comment – if at any stage your gut is telling you that this is all going horribly wrong and isn’t going to deliver value, seriously consider cutting your losses. No matter how bad the pain of your current systems that caused you to go down this path, your current systems may actually be working, and returning to them is better than putting something in place that is totally dysfunctional to save face. Sometimes you need to be brave, and then use what you learned to make a better decision the next time you select and implement an IT solution.
How long will this take?
My experience is that if you do it well, it’s an average of four to eight months until you’re doing business as usual with the new system. If you do it badly, maybe two to three years?