Before I retired, I interviewed or received instructions for numerous patent searches, and frequently the clients were unaware of what they actually wanted.
They might sound sure, but did that mean that they really understood what they needed ? They might ask for a search to see if an invention is patentable, but many do not understand what is essential to achieve that. They usually understand that it has to be new, but may focus on incidentals like saying that the search must be for a certain device which is "light, or portable" -- besides the fact that "light" can mean illumination as well as an absence of weight, it is what the device actually does that should be looked for. The general concept itself may not be of a patentable nature. Someone once asked me if making a known device much larger meant that a monopoly could be secured for just that -- I replied no, it's obvious and hence not permitted. He was amazed.
It is always recommended that someone thinking of applying for a patent consults a patent attorney -- and while some searching may be done before that visit, a search, if possible based on their recommendations, should be carried out after the consultation. There's no harm in the novice having a go so long as it's understood that not finding anything doesn't mean it's not in the patents, it may simply be hard to find.
Even then, the attorney is unlikely to be aware of the linguistic and classification challenges involved. Suppose an inventor says that the invention is for a purring pillow to encourage sleepiness. Can't it be any sort of animal noise ? Indeed, any method of generating sound using the outlined technology ? If a patent application says it's for pillows it means, even if the method is new, it is only protected for pillows. Could the search at least be extended to neck supports or soft toys, for example. In addition, it often turns out that the inventor just has the basic idea and doesn't know how to describe a working example -- in which case a patent is likely to be denied, or will be worthless.
Having said that, there would certainly be no harm in looking for purring pillows to show what sort of concepts were already out there and hence not patentable.
There is for example Purring pillow apparatus by Scott Flack of Illinois, published in 2003. Here is its main illustration.
There were three citations to it, where the examiner found previous patent specifications, including the delightfully titled Austrian specification Katzenschnurr Therapiegerat.
Those citations can be found in the bibliographic format for this patent specification as "Cited documents" (on the left). This application, incidentally, did not progress beyond the published application stage as the only document listed is one starting with 2003 and not a document in say an 8 million number range -- prefixes using the year are applications, it's the other series which have enforceable patent rights in the USA.
In the same format there are three classifications. Only the "cooperative" class is a clickable link, H04R 5/023, which is for stereophonic arrangements for a chair or pillow. It's used by ticking the little box next to it, then (on left) "Find patents" (to find the 838 with that class) or, more usefully, "Copy to search form". This means that additional terms can be added -- such as, say "purring or cat? or animal? or feline" to the second box in the search mask, where both title and summary are searched. ? means that an extra letter, normally an s, can be searched for.
This gives a list of five hits, and at this stage I would study the wording used, to see if the strategy needs to be revised; look at the first description page, to see if it mentions previous work (this is common among private inventors and small companies); and again look at the citations. Any patent specifications I found relevant would have their star ticked to change from black to red so that they appear in the Patents list for more detailed study or listing later on.
These cooperative or CPC classes are not printed on the patent specifications (this is likely to change soon) and will be mainly Western inventions, while the other classes are printed there. CPC as in this case is often a more precise class than the "international" or IPC class, which can also be looked up in the CPC schedules, and is this case are:
H04R 1/02, transducer casings
H04R 5/02, constructional arrangements of loudspeakers
H04R 9/06, loudspeakers
It might be thought a good idea to combine one of those classes (in the IPC box on the search mask) with words such as pillow, cat, etc. Relying just on CPC means that huge amounts of Far East patent documents are not searched. It typically takes 6-12 months for English titles and/ or summaries to be included in the Espacenet database, which is a problem which in theory can be tackled using automatic translation patent databases.
Running a search may be quick, but a variety of searches, each improved or perhaps covering different aspects of the invention, need to be thought through and built. In my career as a patent searcher I never relied on free databases but rather used them as a useful starting point to get an idea of how much was likely to be found, to get a feel for wording that ought to be used, and (if little seemed to exist) to use citations to find older ("cited") or newer ("citing") patent documents referring to the one I knew about. Citations are particularly useful when it is hard to describe how an invention works, as the patent examiners have done much of the work for you. First, though, you have to find a relevant document.
Ideally, of course, the inventor would have been persuaded to augment the search for purring pillows by also searching for how the mechanism worked. Also, the inventor should assist the searcher by explaining terms likely to be used in the industry and hence in relevant patents. I was often struck by how often the novice would wave a piece of electrical apparatus around and say that he had improved it, only to admit he didn't know what it was called when I asked. If he didn't know, how was I supposed to know....
Is all of this easy, or quick ? I would respectfully say no on both counts. I vividly remember the son of a lady client coming in and, in her silent and rather horrified presence, angrily berating me for having the temerity to charge £137 for a search, as he said that patent searching wasn't "rocket science".
He was right. It's often far more complicated.