Skip Links

The rest of how DidTheyReadIt does it

By Mark Gibbs, Network World
June 21, 2004 12:11 AM ET
Gibbs
  • Print

Last week we delved into the intricacies of an interesting system called DidTheyReadIt that lets you track whether people read your HTML-formatted messages, even if you don't enable the "request read receipt" service.

Goof Department: Last week we said messages destined for Hotmail, Yahoo or AOL were special cases and had to be tracked by adding ".didtheyreadit.com" to the address. Turns out we got it wrong, and now everything works fine without the added extension.

But there is a good reason to send mail by appending ".didtheyreadit.com" to messages - it ensures that the message is "burst" to all addressees with a separate ID number for each recipient instead of all being covered by a single ID. And if you use a Web mail service, it is the only way that you can use DidTheyReadIt.

If you don't route your messages via their server, when you examine your online DidTheyReadIt log, you still will see when different users rendered the message. They can be distinguished because they will have different IP addresses and different HTTP referrer strings and browser ID strings. Note that you can have DidTheyReadIt send you an e-mail when each recipient first opens the e-mail and, optionally, on every subsequent read.

Logging could be improved

A few comments: First of all, DidTheyReadIt's logging is far too simple. You can't sort by name, date or any other attribute - the list is simply in a sort of time sequence. The problem is that when an existing record is updated because the message was read again or read by another recipient, the record is updated but stays in the "first created" time sequence. Also, you can't download the log.

And there's the issue that the log will show an entry for yourself when you read a reply that quotes your original message. Not a big problem, but the system should flag the entry.

One interesting thing that the log entries show is read duration. This is determined by using a persistent connection to send the embedded image back to the reader. When the connection is broken, it is safe to assume that the reading has ended.

An option lets you have read times of more than two minutes divided into multiple two-minute reads. This is an odd decision. It would seem more relevant to determine the time it should take to read the text content (fine unless the recipient is a particularly slow reader) and assume that for any duration over that time the recipient simply hasn't closed the message.

How well does it work? Alastair Rampell, the CEO of Rampell Software, publishers of DidThey ReadIt, wrote to us to explain: "I would estimate that nearly 99% of e-mail clients out there render image tags because of the predominance of Outlook, Outlook Express, Hotmail and Yahoo. While we do not have access to the contents of messages sent via our system (and you can verify this by watching the background tracker), we do know how many messages are sent per day, and we also know how many receipts are rendered. Over 90% of e-mails sent during our testing period (with several hundred random users) triggered receipts. . . . If absolutely everyone with whom you communicate is a diehard PINE user, then our software is not useful. But [the majority] of people use HTML-compliant e-mail clients, and even though Outlook 2003 can block external images, a) it doesn't block them for people in your address book or people with whom you regularly correspond, and b) a lot of people turn this off because [it confuses them when] their e-mail from Expedia or PayPal has a bunch of holes in it. That's why we compare the product with caller ID - it will work most of the time but occasionally you will have an 'unavailable caller' or 'private caller.'"

  • Print

Videos

rssRss Feed