Anecdotal evidence may be an oxymoron in scientific terms, but sometimes it’s the only thing standing between you and a pitch-black knowledge void.

Many communications are hard to measure—most organizations aren’t going to conduct a survey of people who received their annual report or a hide a free iPod offer on a key Web page to see how many people read it (though that would be cool—any takers?). That leaves you with selectively gathered or incidental comments.

Relying on this kind of feedback, it must be said, is a bit like relying on your sense of direction instead of a map—you can easily head down the wrong road. Our forays into navigation by anecdote have taught us a few things:

Ask your audience directly. If you’re asking colleagues for feedback, you’re getting a second opinion. If you want evidence, anecdotal or otherwise, you have to ask the people you’re targeting. (And if you do want a second opinion, make sure to tell people what you’re trying to achieve—otherwise, all you’ll gain is insight into your colleagues’ personal tastes.)

Don’t just ask, “What do you think?” The responses you get will be exactly as thoughtful—and useful—as the question. Ask focused questions, such as “Did that X we sent you fill you in on our latest Y?” or “Did the projects we featured in X help you understand what we do?” Construct the questions in advance and make sure they’re not subtly leading people to positive responses.

Write down any comments you receive—solicited or not. Don’t rely on mentally noting anecdotal evidence—you’re likely to lose the original meaning of the comment. And most of us don’t need to indulge our tendency to remember only what we want to hear.

Get a third party to gather the evidence. This could be a firm that does this type of research, but it could also be someone from a different part of your organization. You might not think this is necessary if you’re conducting a small, informal survey, but people are more likely to give honest feedback if they know the person seeking it didn’t have a role in creating the communication in question.

Use with caution. Give anecdotal evidence its proper weight—pay attention to it, but carefully consider the source and how the feedback was generated (see above) before using it to justify a dramatic change in direction or a continuation of business as usual.