Here are some general guidelines on how you can be a better researcher -- and a more effective searcher.
Start with a review
Gather all of the material and information you've collected about your person so far, and make a list of everything you have. The list can be as simple or as elaborate as you want. For tough problems, I use one based on this Genealogy Source Checklist created by Crista Cowan, which she demos in this video from her Barefoot Genealogist playlist.
Once you have your source checklist, create a timeline of the person's life. Extract the information you have from your sources, and put all the events in chronological order, making a note as you go about where the information came from.
If you use Family Group Sheets and pedigree charts, make fresh ones now to make sure you haven't missed anything. You will mine all these materials for ideas on what to look for next.
Before I begin searching, I remember the advice given by Jessica Hopkins, an
Archivist at the US National Archives at Kansas City. In her talk Broke, But Not Out of Luck: Exploring Bankruptcy Records for Genealogy Research presented at the 2015 Virtual Genealogy Fair, she said that in order to find a bankruptcy case, we needed to know the person's name, the place the court case took place, and the time the case happened. This is what Hopkins called the "three-legged stool" -- if you know all three, you have a stable foundation for searching. If you know only one or two of the three, you don't have much to stand on.
The records that are available for us to search are different, depending on the geographical location involved. They will also vary depending on the laws which were in force at the time. See Determining what records are available in a particular locale? for more information.
How general research can help
If I am starting research in an entirely new place, I start a research notebook about that place to keep general notes about it. I include historical maps, information about historical jurisdictions, notes on when counties were created and when towns were incorporated, and other things about the local area that might come into play when I need to search for records. If I am starting a deeper study of a particular type of record found in that area, such as parish records for the Church of England or Non-conformist church records in England, I seek out or build a timeline of important waypoints in history, and look for research guides and finding aids to tell me more about what records I might find, and what information might be found in them.
One valuable resource for doing research about women in the United States is Christina K. Schaefer's The Hidden Half of the Family: A Sourcebook for Women's Genealogy, which includes a timeline for each state showing how changes in the law over time are reflected in the records where we might find women listed by name in historical records.
Records are created for a purpose
A professional genealogist once said in a presentation, "Clerks don't keep records for fun". If we learn about who created the records and why, we can better understand the information they hold, and be better at searching for them too. One thing we can do to better understand the records we use is to find and read the instructions that were given to the people who created them. At ipums.org, a website created by the Minnesota Population Center at the University of Minnesota, you can read the enumerators' instructions for the US Federal Census from 1850 onwards, and get a list of the questions asked about on the census. Census records are usually created to collect statistics about an entire population, and the scholars who use them for population studies often have insights into the information in the census that genealogists miss.
Records have their own genealogy
Once the records are created by their parent agencies, just like children, they can have a separate timeline of their own. Agencies keep records while they are still using them, and at least as long as they are mandated to keep them by law -- after that, if they run out of space, sometimes they dispose of them, and some records deemed to be of historical significance can be transferred to archives or some other repository. There are chains of custody which can be followed. Many records are sent to archives following a "rolling window" rule based on the date they were created. Learning what rules are in effect for a record set can be an important clue about where to find the records (whether you need to look at the original agency, or look for them in an archive).
Learn more about the records
One place to start stepping up your game is to learn more about the records you have already found. If you have anything from a microfilm published by the US National Archives, you can search for a Descriptive Pamphlet (DP) or Reference Report (RR) or other material on archives.gov to learn more. Often the DP for a microcopy publication can be found by browsing all the way back to the beginning of the microfilm roll (or the start of the digital images that make up the 'virtual roll' at an online provider.) These documents can tell you to how the records are arranged, and alert you to records which were filmed out of order, or records which have been lost (things that might explain why you didn't find the person you were looking for when you searched by name).
The beginning of the microfilm roll might also have finding aids on it -- for example, court record books often have their own indexes. If the index was in the same physical book, it may be on the same roll. If a card index was created later, it might be on a separate microfilm. We can use these indexes to find records, too.
While you're re-examining the sources you've already found, you can also make a fresh analysis. Guidebooks like Kory Meyerink's Printed Sources: A Guide to Published Genealogical Records and Elizabeth Shown Mills' Evidence Explained: Citing History Sources from Artifacts to Cyberspace are examples of guides which can show you "where to go to find their sources and how to describe and evaluate them" (the quote is from the publisher's description of EE). Mills' EE website has several QuickLessons, including QuickLesson 17: The Evidence Analysis Process Map, which will help you sort out the different kinds of sources, the kinds of information found in them, and how we can use the information as evidence when answering specific research questions.
One important principle that Mills encourages us to follow is to cite what we use. In a discussion of how to write a citation for a census which was viewed online, she advises a reader:
The most basic issue, where citation is concerned, is Cite what you use. If you use the image, you should cite the image. If you rely on the database, then you should cite the database. Don't mix peas and apples. You'll confuse not only anyone else who uses your research but also yourself at a later time after your recollection of this work has gone cold.
Click through to read the whole discussion so you can see this in its context -- but the idea is sound no matter what we are looking at.
If we come to genealogy from online genealogy, all of the nuts and bolts of how the images are brought to us is hidden from us to make things "easier". We don't understand the process of getting digital images to our screen, which can be something like this:
paper records -> microfilm -> digitized images -> computer indexes -> databases
At any step of this journey, mistakes can be made -- if we aren't aware of the problems, we might fail to find the record we are looking for.
If you've survived so far, make a plan for what you want to find. It's easier to answer a question if you break it down into something small and precise, such as "when did John and Mary get married?" The easiest way is to look for a record that has direct evidence, but what if you can't find a record, or the courthouse burned and the records don't survive? You might have to try a different kind of record --see the article on Burned Counties Research in the FamilySearch Wiki -- or use Inferential Genealogy to answer the question. Make a Research Plan and prepare a Research Log before you search -- it's much easier to keep track of what you are doing if you fill out your log in advance and write down your results as you go.
If you know what specific piece of information you're looking for, you can take advantage of finding aids that will suggest strategies to you, like the research strategy articles, the Record Selection Tables, and the Strategic Research Logs in the FamilySearch Wiki, or articles like Sources of Genealogical Information on RootsWeb. For every site you visit, check their Learning Center, or look for research guides. For research in England, the Society of Genealogists has an series of guides on Getting Started in Genealogy; the National Arvhives [TNA] also has a selection of Research Guides. For the USA, the US National Archives' portal is Start Your Genealogy Research, and the Library of Congress describes its holdings in Reading Rooms like this one for Local History and Genealogy. Local public libraries can have extensive guides to local research and to the material in your own collection. Even if you can't research in person, once you know a resource exists, you can use WorldCat to find it in a library near you.
Searching and Browsing
Now you're ready to search! What are the advantages and disadvantages of doing different searches?
Global Searching is what happens when the big sites invite you to plug in a name, date, and location for your research subject and look for everything. It can be useful if you use it wisely, but if you use it exclusively, you'll miss finding what you want. One useful tip from Crista Cowan at Ancestry is to use Global Search results to look at records in a group by opening each individual search results in a separate browser tab, and rearranging the tabs so that they are in chronological order, so you can walk through the set of records in the order they were created.
Searching an individual database, by going to Ancestry's Card Catalog, Find My Past's A-Z of recordsets, or FamilySearch's Browse All Published Collections usually gives you a wider range of indexed items you can search for, and allows finer controls in how you search. It's also how you can find search results in obscure databases that might not 'bubble up' to the top of the results in a global search, especially if your research subject has a common name. Be aware of pitfalls in each database, such as the problems Crista Cowan demonstrates in her video Some Genealogy Records Have No Names.
For searching some sites, third-party search tools exist, such as Stephen P. Morse's One-Step Webpages. Be sure to read the FAQ to understand how the tools work; you'll learn a lot about the records, too.
If you can't find your person by searching the indexed database, or the database hasn't been indexed yet, you may have to find them by Browsing
the records the old-school way, as if you were actually using the originals and paging through them (or hand-cranking the microfilm reader). This is where all the research you did on the arrangement of the records and their original form will pay off. See Crista Cowan's video Stop Searching. Start Browsing. for an example of how to browse on Ancestry, and use the same skills to browse on FamilySearch, Find My Past, or any other site which allows you to browse.
And -- most important of all -- if you've failed to find your person in your searches, start writing things down. If structured research logs aren't helping, then write it out in a research journal. Write down exactly where you searched, how you searched, and what the results were. Think about reasons why your search may have been unsuccessful. Did you choose the wrong wildcards for the site you were searching on? Did you search over and over again for a marriage that might have taken place in 1920, using an index for marriages in that state which only goes up to 1910?
It seems foolish to suggest that you write out "I didn't find John and Mary's marriage in the Massachusetts State Archives' Vital Records Search -- they may have married in 1920, and the search stops in 1910" -- but if that's what it takes to remind you not to search there again and again, and move on to finding the information a different way, then do what works.
Writing out notes for yourself -- thinking about how to write up a focused question for Genealogy.SE -- actually writing things down, in a form you can go back to again, allows you to leave a snapshot of what you were thinking at the time, that you can come back to later after you've improved your skills. It allows you to take a class or webinar and then come back to your research and put the new ideas into practice. If you can't find something, try to puzzle out why you haven't -- and write it down.