Freelancing Stack Exchange is a question and answer site for self-employed and freelance workers. It only takes a minute to sign up.
Sign up to join this community
If I have a client who is asking me to solve an automation problem that involves integrating different software solutions, like a payment gateway and a form builder tool, for instance, and we don't know if the integration will work the way he wants it to, how should we proceed in terms of estimating the costs?
Basically, it's possible that error handling may be problematic and either would not be possible or would involve more work to find a solution.
When I have faced that scenario in the past, I have performed a small (10-20) hour engagement to define the requirements and design a solution.
The document that I put together then serves as a road map for the next phase, which could be performed by me or by another resource.
I've worked on projects with unknown variables, and they tend to be a big pain in the rear. What has worked for me though...
Setup a time to talk with the client, and get information about what he wants done. This is much easier if he does not have a preference to the software or tools, and is willing to have everything happen from scratch. But you need somewhere to start with the project. Include any obscure feature that they want, especially if you are not building the project yourself.
For product recommendations, you will need to do research. This time needs to be billed somewhat, as very few customers will pay a full rate for research. If you can discuss this beforehand, you may get away with it, but in my experience, no one would pay more then a couple hours of labour for research.
When I talk research, I also mean to actually talk to the companies, and their technical team, to ensure that product X from Company Y will work with their Product Z. Get everything in an email format so there is a paper trail, and it includes any weird specifications that you may lose track of in just a phone conversation. You need to be on the ball!
You will also need to check all the system requirements, license requirements, and any external factors needed (i.e. a separate LAN line or phone line). If the customer's machines need upgrading, that will need to be done before you recommend Product X and Z, and you will need to discuss who will cover the bill: part of your quote, or get another company to do that.
Find out if there are trial versions of all the software available. You will need to get copies of the software, and ensure everything works together! As well, installing it on your own test machine will help you find out if there are weird setup instructions, or special requirements not mentioned elsewhere. I say to use trial versions as you will usually need a miracle to return opened software, and you should not have to eat the costs of that at this point. You haven't got the job yet, so you should not be out a few more thousand dollars!
Once you have the final estimate, and every last feature is defined, write it up and have the client sign it! This will then go with your Scope of Work, and your contract. Include a list of the companies that will help support features A, B and C, and who will support features D, E and F. If you are covering all the support, ensure your costs are built in there, or the client understands your hourly rate for fixing products.
The biggest part of doing your homework though is to keep poking and prodding. One great thing about working with certain Windows programs is that many developers have public APIs available, so many other products will work, right out of the box. Not every program is like that though, so care must be taken to ensure all the requirements are met, and everything works as expected.
This is a pretty hard scenario but this is what I do in this situation:
If they aren't happy with that sort of arrangement then depending on where they are based you can offer to work on-site for the first few weeks, some people like this because they can see you doing the work, and they are more likely to trust your time estimates.
