I have been an avid cruciverbalist – person skillful in solving crossword puzzles – for the last decade.
Though in my younger years the NYT crossword was impenetrable, only the easy Monday crossword presenting any chance of consistent completion, with the introduction of the NYT crossword app I was able to get enough reps to crack the previously enigmatic clue conventions.
For the past year or two, I’ve been able to solve almost every daily NYT crossword, from the straightforward Mondays to the tricky Thursdays right through to the Saturdays, the most difficult puzzle of the week.
And, I’ll admit this in present company: sometimes I cheat. When there’s a particularly perplexing clue that I can’t parse, I have no shame in searching Google to find a bit more direction.
One recent night while searching, it gave me pause. Others must be doing the same thing. I wondered if I could use tools from my work in SEO and digital marketing to observe any kind of pattern or trends around these searches. I got a little excited seeing my home life and work life cross, and decided to investigate the search implications further.
It turns out I’m not alone. Using Google Trends data, you can easily see how search engine data correlates with people searching for additional clue information – or, in less forgiving parlance, cheating – for the NYT crossword.
We see a huge spike in searches for obscure subjects referenced in the crossword for a day or two after the publication of a crossword, only for the trend data to fall back to obscurity.
Here are some quick examples of search spikes, which we can only reasonably conclude are tied to searches for answers to NYT Crossword clues.
On Tuesday, November 19th, the 34-across clue asked the solver to name a “* cartoon billionaire.” The asterisk, tied to the puzzle theme, held the additional information, revealed in 45-down, that the character in question wore a top hat.
The answer was “Scrooge McDuck.” For a healthy volume of solvers, they took to Google to find the answer instead. Note the spike in searches for “cartoon billionaire,” which aligns with the publication of the November 19th crossword.
When I don’t have the answer to a clue, I find I’m more likely to…seek additional information…if I’m given very Googleable content like names and years. Though I knew the answer to Saturday, November 9th’s 39-across clue “1981 Rick James hit that starts ‘She’s a very kinky girl,” these keywords saw a spike around the publication date.
Check out the trend data for “kinky girl,” which spikes on November 9th:
In the same vein of very Googleable answers, check out 1-across (a terrible start if you don’t know the answer!) “Von Trapp daughter in ‘The Sound of Music,” which was published on November 1st.
You’ll notice the trend line is pretty modest for some of these terms. “Von Trapp daughter” gets negligible searches, so even a handful of new searches will cause a spike. Yet even for a keyword with sustained popularity, we see the effect of NYT crossword searches.
Take “zodiac ram,” both the clue for 27-down on Tuesday, November 12th and a keyword with an average of 1,300 monthly searches in the United States. It also peaked on November 12th.
It is only when the keyword is so generic that trend data offers no insights. Consider “flip out,” the 32-down clue on Tuesday, November 12th and a keyword with an average of 27,100 monthly searches in the US. Trend data shows no correlation with the NYT crossword release.
But return to the comfort zone of rare terms, and the NYT crossword impact is readily apparent. Guess the crossword release dates for these trend lines:
Though the search volume for these keywords is low, there is a whole subset of sites that are capitalizing on this trend. For instance, the keyword “bilingual muppet” brings up the following domains as the top results:
As you would expect, these sites are loaded with ads but offer useful tools like search boxes where you can enter partial answers (ROSALI??) and be given a likely correct answer based on its large database of crossword clues and answers. Crossword clues and answers are often repeated. For instance, the common crossword answer “ELLA” (often clued with “Jazzy Fitzgerald” or “Singer Fitzgerald”) has been used 204 times since 1993, the beginning of the Shortz Era named for editor Will Shortz. Name any word or phrase: Ottawa (39 times), Taylor Swift (4 times), Betelgeuse (2 times), and there’s a chance it exists in the NYT clue or answer database.
What is also surprising is how quickly these sites are able to update their databases as first-time clues and answers are published.
For example, the Friday, November 22nd puzzle was released online on Thursday, November 21st at 10 PM (it is standard for the online version to be released the night before). This puzzle featured a vocabulary debut: SITUATIONSHIP as the answer to “Romantic gray area.” This was the first time this neologism had ever appeared in a NYT crossword, so it wouldn’t exist in the pre-established database.
Yet within the same hour of the crossword’s publication, a search for “Romantic gray area” turned up a bunch of listings to crossword answer sites, all listing the correct answer.
Whether these sites are capitalizing on early access to the crossword or simply have lightning-quick fan contributors, the dedication to providing answers to cheaters as quickly as possible is readily apparent.
And of course, the Google Trend data shows the spike in searches which will lead to hits on these sites:
For so many of us, Google is the first stop to solve all the world’s solvable problems. Whether you want to check the weather, buy concert tickets, or learn how to make ratatouille, Google is probably your first stop.
Google also plays a central role in our hobbies. An athlete looks up new bodyweight exercises; a musician looks up charts for the mixolydian scale; a gamer looks up a walkthrough of a particularly challenging boss in Dark Souls; and, clearly, crossword enthusiasts look up the answers to solve puzzles.
Thanks to the time-sensitive nature of the NYT crossword puzzle, we can isolate these particular searches to see how Google, and the advertising-fuelled answer sites it lists, help solvers solve a little more successfully than they would on their own. Myself included.
P.S. And for the record, two Google searches were used in my completion of the Friday, November 22nd puzzle. The first, to confirm that “‘Ghostbusters’ co-star Rick” was indeed MORANIS. For the second, a 4-letter clue for “German automaker.” AUDI wasn’t working and BENZ didn’t seem right. A Wikipedia list pointed me in the direction of OPEL. Cheating? Creatively using available resources?