I spent yesterday morning estimating a client's marketing uplifts. Well, when I say estimating, guessing would probably be more accurate.
There's nothing especially wrong with that. If you haven't got any idea what your marketing ROI might be, an analyst can usually put you in the right ballpark with a small amount of effort and some 'creative' use of sales data.
There are two ways of going about guessing your marketing returns. One is right and the other one feels right, but leads to totally unfeasible levels of sales almost every time.
Lets start with the wrong way. You know those people who get their business plan laughed off Dragon's Den? The ones who say "I've made a product for the pet care market and the UK pet care market is worth £1.5bn per year. If we just get 0.5% of that..." Then Peter Jones cuts them off and asks how many kitty litter boxes they sold last month. And the answer is less than ten.
When you say "My competitors combined, have an 80% share of market. If I can just get x% of that" you're making exactly the same mistake. Or The other one you hear a lot is "there are loads of people who only buy my product very occasionally. If I could just turn 10% of them into regular users..."
The right way to do it is start from where your sales are now, plus the sort of marketing spikes you can see in the past, with some sensible adjustments bolted on. This gives a much smaller number than the first way, but is a lot closer to the truth and won't get you laughed at on the BBC...
All those people who don't buy your product now are not buying it for a reason. One ad isn't suddenly going to make a lot of them change their minds.