If you fire a bunch of bullets at the side of a barn, and then walk over and draw a bullseye around the tightest cluster, you can claim you're a sharpshooter. That's where this fallacy gets its name. The Texas Sharpshooter fallacy is all about cherry-picking data after the fact to make it look like a meaningful pattern was there all along.

It happens constantly in marketing. A product might claim, "Four out of five dentists recommend it!" but they forget to mention they surveyed twenty dentists and only included the four who gave the answer they liked. Or they'll say, "In the cities where we have the most customers, people report feeling healthier," without mentioning that they only looked at cities where health stats already trended high. That's drawing the target around the arrows after they've landed.

In business, this can show up when a manager presents a performance chart that highlights one strong month and glosses over the previous eleven. Or when a company celebrates a rise in sales by ignoring the fact that profits are down and customer complaints are up. Data can always tell a story - the trick is whether you're telling the whole story or just the part that fits your narrative.

You see this with PC buying advice all the time. Someone insists, "AMD is better than Intel," and then links to a handful of benchmarks where their favorite chip comes out on top. But they conveniently ignore other tests where Intel leads, or real-world usage where the difference is barely noticeable. Cherry-picking only the data that proves your side isn't analysis - it's just marketing dressed up as expertise.

And of course, it shows up in Microsoft Access forums all the time. Someone will claim, "Access corrupts your data," and then cite a few horror stories they found on Reddit. But they ignore the fact that those examples almost always involve poor network setups, improperly split databases, or users editing live data over Wi-Fi. When you consider the sheer number of Microsoft Access installations around the world - over 94,000 documented companies (1) across industries and likely many more small businesses and internal tools - that's a staggering number of successful deployments. Sure, you're going to have a handful of horror stories. That happens with any platform. But it's not statistically significant. I've dealt with literally hundreds of companies using Access over my 20 years of consulting, and hundreds of thousands of Access users through my YouTube videos and courses. The number of people who experience serious, irrecoverable corruption due to Access itself - not bad user behavior - is a drop in the bucket. I can count them on one hand.

I once worked with a client who was convinced his Access database was "changing numbers on its own." He pointed to a sales report where two records from last month didn't match what he printed out a week earlier. He was sure the system had been hacked or the data was corrupted. But as I dug in, it became clear he was looking at two different versions of the same query - one grouped by customer, the other by product - and he hadn't realized they summarized the data differently. Instead of reviewing the entire report logic or checking how the queries were built, he fixated on those two mismatched numbers and built a whole theory around them. He was drawing a bullseye around the bullet holes.

When you're consulting in Access or any data-driven system, this happens more often than you'd like. Clients latch onto a handful of data points, decide something must be wrong, and ignore every log, audit, and validation rule that says otherwise. The best way to handle it? Stay calm. Ask them to walk you through their exact steps. Rebuild their path from input to output. And then show them how the misunderstanding happened. Don't fight their theory head-on. Show the map, not the monster.

In science, we try to avoid this by setting our criteria before we gather data. You don't run an experiment, then go poking through the results looking for some correlation you can publish. That's how you end up with bizarre headlines like, "People who eat more cheese win more Nobel Prizes." (Yes, that correlation exists. No, it's not causal.)

This fallacy shows up all the time in climate debates. Someone sees a single cold snap in Texas and suddenly declares global warming a hoax. Or they point to one year where Arctic sea ice rebounded slightly and use it as proof that everything's fine. But that's classic sharpshooter logic - drawing a bullseye around the one dot that fits your narrative while ignoring the broader pattern. Climate science is about trends, not outliers. If the ten-year curve says the planet's warming and the ice is shrinking, you can't toss that out just because you had to scrape your windshield in April.

Politicians love this fallacy. They might highlight a single favorable poll and ignore the rest, or boast about a district's job growth while ignoring wage stagnation or inflation. It's all about cherry-picking the data that makes them look good and sweeping the rest under the rug.

You even see this in moral or philosophical debates. Someone points to a handful of ancient verses or teachings and says, "See? This proves my point." But they're cherry-picking from thousands of years of contradictory texts, ignoring the parts that don't fit. It's like drawing a circle around the handful of quotes that support your worldview and pretending the rest never existed. Human history is messy. Wisdom isn't always consistent. You have to take in the whole picture, not just the parts that flatter your position.

In Star Trek: Voyager, there's a perfect example in the episode The Voyager Conspiracy. Seven of Nine starts absorbing massive amounts of ship data, far beyond what she was designed to handle. She begins to see connections between unrelated events and constructs elaborate theories out of nothing. First, she believes Janeway is hiding a secret agenda. Then she flips and accuses Chakotay of conspiring with the Maquis. In both cases, she cherry-picks logs and patterns that fit her theory while ignoring everything that doesn't. It's a textbook case of forcing a conclusion and then drawing the evidence around it. She wasn't uncovering truth. She was falling victim to the fallacy. It's not that the data pointed to a conspiracy. She wanted a conspiracy, so she found one.

The bottom line? If you're only looking at the hits and ignoring the misses, you're not analyzing data - you're decorating it. Real insight requires looking at the entire picture, not just the convenient parts.

LLAP
RR

(1) The usage statistic for Microsoft Access comes from Enlyft, which tracks technology adoption across industries. While Enlyft's data isn't exhaustive, it's generally considered a credible source for estimating tech adoption - especially among enterprise tools. Think of it as a well-informed sample, not a full census. That said, the 94,000 figure is likely conservative. Many Access installations are internal and never appear on public websites. Thousands of small businesses and independent developers use Access without any visible IT footprint. And Microsoft no longer publishes official usage stats for Access, which makes it even harder to capture the full picture.