I am researching my competitors before getting into SAAS based software market. For one of the competitors total monthly unique visitors to be 3000. I am calculating roughly 50% of them are non-customers coming from search engines etc and rest are customers. Does this make sense?
this is very very hard to guess, but people are right in the comments that compete.com or even Comscore are not very precise. I had Comscore for a past company I worked at and it was about 50% off the real number. Remember that there is NO way for them to track your traffic so they base that on random panels of people. Comscore switched to what they call "Hybrid" which mean you give them your log files and they combine panel data with your logs which makes it much more accurate, but not everyone is using Hybrid with them. I'm guessing only the big guys do it (we were one of them).
In any case, I feel that 50% is much to high. You would be surprised at how much useless traffic most sites get. Especially when you are talking 3M uniques (I'm guessing you 3000 are in thousand otherwise it would be really really low - untrackable). The first thing to think about is the "Abandonment Rate". A LOT of people land on the site and leave right away. This can be somewhere between 30% and 90% so off of the 3000, you can shave off somewhere around half of the people that didn't really want to get there. Out of the rest, you'll get people that are just looking around, somewhere between 30% and 80% let's say, so you are left with another half off there.. and out of the rest, if your competitor has a Freemimum model, you can easily count that 90% off the users are free.. You are left with a very low number.
In any case, when you are doing your financial forecast ALWAYS use a sensitivity table where you plug in your assumptions. play with the assumption and determine which one is really import => if one variable changes by 1%, how much of an impact does it make to your revenue? figure all that out and it will give you an idea of how important those assumptions are and how much effort you need to put into the accuracy of those same assumption. does it really matter if your competitor has 50% of paid customer or only 5%? How does that change your model? etc..
Have fun. hope this helps.
I'm not in the business, but I own a website, so take my answer with a grain of salt.
I'm guessing that more than half customers came from a search engine and were searching for something else. Of the rest, most are probably just checking out. Actual customers are probably just a fraction of visitors. But that of course depends on which market you're in.
I think you need another proxy for market size. As others have explained here, Alexa, Compete.com and the rest are massively wrong most of the time. By several orders of magnitude. Even if you are able to accurately estimate their traffic, you have very little idea as to what that breakdown might be in terms of customers, prospectives etc.
The SAAS websites might not serve their service over the same domain. If the competitors' services are accessed over a different domain to their main marketing website, then any toolbar-based traffic estimator is going to completely miss that traffic.
Can you access your competitors company accounts? Depending on where you are in the world you may be able to access the annual accounts filed by your competitor. This would be rather nice to have...
Going back to your 50%... my feeling is that is way to high. Reduce by 10x at least.