Always take into consideration what your customers want, even when building "what they really need", because it will give you insight into how best to implement it.

On whether to "give them what they want", it comes down to what % of your customers want something. If 99% of your customers want something, it would probably be best to give it to them. If 5% of your customers want it, don't bother.

If you're just reading posts to a "things I want" forum, then what you may be seeing is the opinion of a loud minority, which will make it look like 99% want something, when really only 2% do.

To accurately assess the % you could actively send out an email poll to all your users asking which of several options they would most want (allow them to choose 1). Embed the poll in the email itself to make it as easy as possible to respond to. Also, if possible embed in the email a screenshot of what each new feature would look like, so that it will be easier for them to see your side of the story about why you think a feature would be "what they actually need". Just have your designer make the screenshots, you don't need to actually have them be programmed and functional.

If you want to discuss this in relation to the unique details of the actual features you're thinking about implementing, let me know,

best,

Lee

I would suggest market research to ensure you have the most accurate data.
When we launch an update to our existing products or even a new line at Vital Consulting, I always make it a point to sell the experience and influence the customer of what they want.

Help them discover for themselves what their needs are and turn them into wants.

There are a few conflicting statements in your question that make this a little tricky to answer - but I am gonna give it a go anyway.

Firstly, it seems like when you're saying "what they really need" perhaps you have strong opinions on what they need and are frustrated that they don't see things the same way as you? If this is the case, you are definitely going to be facing the uphill battle that you reference.

My suggestion is to talk to your customers (actual face to face, or telephone calls) and find the trends in the features or specifications they are requesting. If you are already established (which I am taking to mean - you have loyal, paying customers and are financially stable) then take an 80/20 approach to improving your product.

Spend 80% of your time/resources by adding the feature requests that you know people are asking for and likely to pay for and spend 20% of your time/resources on those experimental features that you think people need even if they don't realise it yet.

Also, this will probably keep you motivated to continuously test and improve your own product because you'll have a 20% "side project" budget. Those are always fun.