After stating my general policy that a search is always recommended, I always ask potential patent clients who ask about a search. The answer to this question lies in answers to several other questions.
THERE ARE THREE TYPES OF PATENT SEARCHES
The first is a patentability search is for seeking out and evaluating references to see if their teaching would prevent an applicant from receiving a patent. These searches are typically brief, performed for under $1000.00 including the attorney opinion, but what is their value? Are they statistically relevant?
The total body of information which can to prevent a patent from issuing, or invalidate an issued patent, includes all of the publications in the world, going completely back in time, in all languages on planet earth. Books, magazines, theses, technical manuals are a small part of the total body of information, and includes everything from an undergraduate paper in Chinese in a library at the University of Beijing, to a pamphlet in Japanese in the library of the University of Tokyo. Thus, the number and types of places which can be searched are vast, and to cover even a significant portion of them would be quite costly.
A search can only affirmatively affect the patent process by producing a blocking reference; a search uncovering no close references does not mean that none exist. Given the large body of information available and the relatively small amount of it searched, the lack of a blocking reference or references may have only a small statistical significance.
If the object of a patentability search is to find "if the invention is known" the analogy is much like making a decision that no tiger is present in a building with 1000 doors based upon looking behind ONE door and finding no tiger.
A patentability opinion is, and must practically be, limited to the results of the search. The opinion correlates with the world's body of information only as the information ACTUALLY SEARCHED becomes significantly large. An expensive, extensive search will make the opinion more highly correlate to the real world. But searching all the worlds patent offices (or only Europe, Japan, & the U.S.), and all of the great universities, etc., is considered by some people to be prohibitively expensive. But if a "brief" search is in the right price range, why bother to get one, if it misses so much of the available body of information?
Previously I considered a "correlation search" trying to match the Examiner's likely results in the patent search room. With todays many internet options, I'm not sure if a search in the PTO manual files would correlate well. Examiners are increasingly eclectic in thier searching.
A second type of patent-related search further qualifies the limitations concerning the first, or patentability search. An infringement search typically seeks out and analyzes only U.S. Patents which are currently valid, the search based upon infringement of a particular device. The cost for this search is usually between $2000 to $10,000 depending on how much search time is directed to be spent. This search attempts to gauge a producer's liability for making, using or selling (producing) the potentially infringing article. This type of search has a short sequence of patent discovery followed by classification to threatening and benign by the attorney, the threatening patents then subject to further analysis by study of their file histories to better understand the scope of the claims. The liability is always open because no search is perfect, and besides, the patent which may be infringed by producing the particular article might issue the day after the search, and cannot be currently searched since patent applications are confidential. Many inventors erroneously believe that a non-negative result on a patentability search gives a green light to non-infringement, but as has been shown, even a non-negative result on an infringement search doesn't foreclose liability for producing the article. Producers reduce their liability in this area by purchasing a commercial liability policy with a patent protection rider.
If one or more patents from the infringement search remain as a threat to the article to be produced, the third of the patent-related searches, a validity search, may be used to try to find references which significantly pre-date the filing of the threatening patent, and which are close enough to the threatening claims to render them invalid. The validity search is usually time interactive since it is usually not desirable to search everywhere in the world at once, but to proceed with the search in each resource sequentially. A validity search is also what a producer considers using when faced with a patent holder demanding a royalty. The sum-certain royalty with its finality and closure on one hand is compared against the cost of a plan of a sequence of validity searches. After each negative outcome in the validity search, an evaluation of the pay-or-search-further choice is made. From the producer point of view, each additional amount of money spent on searching could yield nothing, leaving the original demanded royalty still to be paid. Or the search can result in a single reference invalidating the patent outright. In between, references might be found which weaken or eliminate only some claims. The extent to which a producer will search to invalidate a patent depends upon the value of the product and the severity of the patent holders demands. If the patent owner demands are small, it may cost less to pay the patent holder than to initiate searching. If the patent owner demands are high, and the product value is low, the producer may simply abandon production. But if the product value and patent owner demands are high, the producer will be more willing to spend significant sums for a validity search. Thus, much depends upon the magnitude of the patent owner demands.
Which brings us back to the patentability search, the first patent-related search. If the value of the invention and demands on a licensee will be high, it may benefit the patent applicant to spend significantly on a thorough search to determine the true value of the invention, to bolster his knowledge of the value of his invention and help his negotiation process.
But where the invention value and the applicant's demands are similarly not high, the perceived need for a search may be more driven by attempting to lower the risks in the relatively low cost expenditures for applying for and prosecuting a patent application. But do the economics of using a search in this way make sense?
A search will add cost to the patent process both due to the cost of the search & opinion, as well as the cost of preparing and filing an information disclosure statement to disclose the search results, if the results fail to be negative.
A producer will consider the cost of a patent application as just another small expense in protecting a product, and thus not consider a search as a qualifier for spending to apply for a patent, due to the value in being patent pending at least through the prosecution phase. Non-producer inventors are more likely to consider a search as a test for deciding if risk of expenditure for a patent application is warranted. Lets look at the economics.
Since a design patent application usually costs less than $1800.00, the cost of a "brief" search and information disclosure statement (if the search results fail to be negative, say $1800.00) can easily exceed the cost of applying for the design patent, is it really worth doing a search based upon cost alone?
A small utility patent application with 23 pages of specification and 2 sheets of drawings is expected to cost less than $5000 to file. Compared to an $1350 cost of the search and disclosure statement, a ratio of about 4:1 still does not seem particularly advantageous as a decision mechanism, especially since the "brief" search is limited and has a low probability of finding the best references. Where the patent application is much more complex, such as 70 pages and 15 sheets of drawings costing about $16,000 and having a ratio of about 12:1 doing a "brief" search upon which a decision to continue the patent process will be based, is better than the 3:1 case, but still not a reliable mechanism to decide whether to proceed. Even where references from a search can be discovered for free, it still takes time to analyze them. Further, they have to be cited on the filed patent, and further time and cost is expended for the information disclosure statement.
In conclusion, where the invention is valuable, the inventor should have an expanded patentability search performed in order to properly evaluate the highest marginal value of his ideas, since a potential licensee will order his own expanded validity search to test the patent. For a non-producer inventor whose invention is of low or questionable value, it makes less economic sense to use a brief search on which to base a decision to move forward to obtain a patent, because it may be statistically un-correlated to the probability of success or failure in getting the patent allowed. Searches are always recommended, but the applicant should evaluate the economic efficiency versus the statistical relevance.