Training -- though sometimes stressful -- is one of the best parts of my job, largely because of what I get to learn from people with completely different outlooks who pose incisive questions I would never think to ask. On the orientation-to-Salsa webinars we've been running for current clients, I got a great question the other day about running reports to help campaign managers identify possible mistyped e-mail addresses so that they could manually correct them in the headquarters while the relationship was still retrievable. A perfect occasion for both our custom report builder and a forehead-smacking "why didn't I think of that?"

What Typos Do People Actually Make?

I expect there are some much more comprehensive studies of this out there somewhere (please chime in if you know about one). But as a starting point, both for this project and for the inherent interest, I decided to run some searches to see just how usual some common typos might be. And the survey says... Note that the Google spreadsheet linked above is a search on our main data core of approximately 10.6 million, so it's only 60 to 70% of the actual entire DIA community. Note also that I only ran the searches I could think of. I'm confident I've omitted something blindingly obvious -- if anyone has something to suggest, I'll be happy to check it. And note that each query was run in isolation, so someone who signed up as test2yaho,co would count towards at least four different searches. Just the sort of folk you want to be sure are on your list. Allowing for that caveat, and assuming that's the exception rather than the rule, the raw sum of these searches is around 50,000 -- nearly one half of one percent of the data in question. That suggests some real value in attending to these typos. They're separated into general problems and domain-specific ones for the top four mail domains (Yahoo, Hotmail, AOL and Gmail). When reading it, the % sign is a wild card, although I may not have been completely precise in each case about the exact query I ran, it should give an idea. Want to skip the spreadsheet? Know that by far the most common typo is using a comma (,) instead of a period (.).

Cleaning Up After Klutzy Fingers

It turns out that our intrepid developer Erin has actually already built a script that does some of this automatically on the input side -- for instance, an address such as %2%.% without an @ (the supporter has likely typed "2" in place of "@", failing to hold down the "shift" key) is automatically rewritten. As an additional failsafe, I've also built many of the most straightforward typos above into a report in Salsa called "Possibly Mistyped Email Addresses" and tagged "supporter" and "email".