If a private company provides a grant for some research to be carried out but is hoping for a particular outcome from the research in order to take certain business decisions, should this be described as a conflict of interest?
I don't see this as a conflict of interest, but the issue itself is deserving of careful attention from both a contracts perspective and a scientific perspective. I would advise against any such funding source that puts restrictions on the publication of findings. In particular, you state that the business is "hoping for a particular outcome." Are you still allowed to publish results that are contrary to their expectations -- or, does the contract specifically state they have authority over what can and cannot be published. You will want your grants management officers to carefully scrutinize the contract.
Science is really about answering questions through an open, replicable and systematic process. It is not about providing evidence for an expected outcome. I think this type of funding will be more problematic than helpful for a scientific career.
A company exists to make money. They fund a research in the hope that the outcome will be beneficial for them (as a student works on a project in the hope that there will be good results out of it). There is nothing intrinsically bad there.
The problem may be with the rest of the conditions. But it will be perfectly fine if: (a) They guarantee you are free to publish whatever you consider and (b) they don't force you to get "good results".
Yes, it is a clear conflict of interests:
a situation occurring when an individual or organization is involved in multiple interests, one of which could possibly corrupt the motivation.
It does not disqualify your research (as long you make it with full scrutiny, etc; see other answers), but for the sake of scientific honesty you need to mention it (as it can possibly alter your motivation, e.g. to get further grants from the same company or to condition probability of publishing on the outcome).
I saw such notes, e.g. that a given scientist works for a particular drug company.
Yours and their interests do conflict but this is not how we usually use the work "conflict of interest." Say if your company actually makes cat food, and then you later got recruited as a consultant on setting up a safe cat food policy, your decisions made in the policy task force can be potentially affected by your financial involvement with the company. In that case, you will need to declare to the policy task force that you have a conflict of interest. In most cases, conflict of interest requires more than two parties, or one of the parties having two or more roles.
In your case it sounds more like a threatening to academic freedom and scientific integrity. Some preventive measures should be done now before things turn ugly. Here are some things that I have been doing in various projects, you may consider following any of them.
Set up a "statement of understanding." It's a binding contract describing who should have the interpretive rights to the data, and how to safeguard everyone's interest. There isn't a template for that, you'll need to lay out items you cannot risk to lose (e.g. your reputation, your institute's reputation, etc.) and contrast with the company's intentions. Talk to your institute's PR or legal representatives for a consultation.
Highlight your value as the neutral middle man. In that sense, express to them firmly that you would like to explore the quality of the product without any prejudice. And explain to them that if their product is substandard, it's better to know now than later when found out by consumer interest group.
Agree upon ALL the protocols, big or small, prior to any analysis. The turmoil I got into mostly about reanalyzing data. Some people just don't give up... and if you look at the same thing for enough iterations, good findings can come out just due to chance. Follow the protocols strictly, and meet to agree again on any modification.
Keep all documents, e-mails, lab books, etc. and keep copies of them.