Question of the Month: When you're choosing components for a new project, what influences your selection the most?

Although I selected "Detailed datasheets and documentation", there are several criteria that use when making a selection (several of your selection).  Cost is an important factor, but Distributor, Manufacture and package type are also important considerations. 

I felt cost is too obvious.

Although I chose availability, it is not as simple as that, when looking for components I find the ideal candidates through data sheets etc. but if it is unobtanium then that's useless or if its a shady / potentially unreliable source then again, useless so it comes down to finding the right component that can also support the life of the product, this may not be the "Best" component

cstanton Reliability.

I tend to choose components based on the following criteria - in order:

1. What I have in stock
2. Which component has the right features (including cost) for the application
3. What is in my CAD library
4. Which supplier has it in stock
5. Which supplier has the lowest cost (including shipping cost)
6. Which supplier has the shortest delivery
7. Which brand I have used successfully before
8. Which brand has the best reputation
9. Which supplier has the best reputation

Sometimes I will buy everything from one place for convenience, even if one part is slightly cheaper elsewhere.

If everything is essentially equal, I would buy from Avnet/Newark/Farnell just to support element14.

Documentation has dictated which supplier several times, even when the cost was a bit higher, I chose the supplier with the better documentation.

I tend to do research before making component decisions, so datasheets and documentation are important at that stage, as is a good parametric search engine.

Sometimes I will buy something just because it is a great value, even though this doesn't always end up being useful.

Sometimes (quite often) I will buy more than I need to take advantage of price breaks. Sometimes this takes a long time to pay off...

Sometimes I buy extra because I went to the trouble of putting it into mu CAD library, which also means I will likely order more PCBs than I need.

Sometimes I will up-sell myself (get a better component than I need) to reduce future learning curves.

Sometimes I will down-sell myself to eliminate a learning curve or use up stock.

Cost only comes into play after you find a selection of components that do the job.

Depends on your design target. Fast works approach doesn't worry about reliability in the first alphas or betas as the design is likely to change specs anyway. Certainly going to market reliability is important for medical equipment,  but maybe not so much in a childs toy.....just as an example for discussion 

robogary Of course, selection factors vary wildly depending on your work.

Hard to know which order to rank, to decide which one to vote for, but I wanted to touch on the "community recommendation and reviews" - more often than not, the products that are reviewed, or the manufacturer blogs regarding new products, have been extremely useful. New components have new features which I often only initially learn from blogs or reviews. For instance, I'm having an interesting time trying to develop a gesture-controlled switch using a part that a manufacturer blogged about on element14. And I strongly considered Infineon parts for a load switch based on a blog of theirs (and I did order some of their parts, but I've not narrowed it down yet, since I'm also looking at NXP).

We should be allowed to rank these rather than just being allowed select just one option, as all apply to varying degrees.