|
useit.com |
| Search |
Jakob Nielsen's Alertbox, August 20, 2000:
Summary:
Mailing list content must be ultra-short. Provide separate email addresses for subscribing and unsubscribing and include info on how to get off in every mailing list message. Improved usability increased subscriptions by 128% in one case study.
Email lists are an e-marketers dream: mailing lists provide a highly targeted way of reaching people; email doesn't require you to wait until the customer remembers to go visit your site.
Mailing lists allow you to extend the footprint of your website. In the literal sense (get space in the user's inbox and not just in the browser). And in the more interesting metaphorical sense: More services become possible when you can reach out to users and provide them with time-dependent information.
Just remember the push fiasco: it is not the goal to lay claim to ever-increasing amounts of the users time; prompt them just enough to be useful but not so much that the email becomes a burden. Users will unsubscribe faster than you can say "information overload."
Some companies send out very long emails that constitute complete newsletters. I prefer keeping the email very short, restricting it to headlines and summaries with links to more extensive content on the website. Two reasons:
Bad usability. No interaction. Plenty of cause for errors.
If everything went well, the user would receive instructions in the use of the mailing list system. Do users read instructions? Worse, do users save instructions until the day they need them? No, that's why all these old list systems are flooded with mail asking "how do I get off?"
It is better to make the email system itself serve as the user interface: provide a special email address for each of the main commands: subscribe, unsubscribe, post discussion group message, etc. Keep the list short and there is a chance that users will understand it.
Any more complicated commands should be relegated to a web-based interface. Hopefully even the most advanced mailing list system will be simple enough to be explained on a Web form with a few buttons and type-in fields.
Sure, some users will overlook the box and will not uncheck it. That does not mean that they will welcome your newsletter. On the contrary, they will detest your company for sending them spam. Even worse, those users who do notice the little box will lose trust in your site because they will feel that the site is trying to trick them.
Much better to leave the box unchecked and only collect those users who are eager to get your material.
The most trust-destroying design I have seen was employed by Boo.com: they had the standard little checkbox, but the fine print said to check the box if you did not want to get their mailings. As we all know, users don't read instructions very closely, so the average user would simply leave the box unchecked, believing that he or she had avoided the newsletter. When the mailings started arriving anyway, users might be very upset about the presumed violation of their instructions. Luckily Boo died because of its many sins against usability, so this "opposite design" did not hurt very many people.
To increase the number of users who subscribe, don't just talk about "valuable offers". Instead, provide explicit information about:
It is nice to provide a feature for getting each posting as an individual message for those users who want to follow the discussions in real time.
Discussion groups should preferably be moderated to prevent spam from making it onto the list and to enforce a minimal set of standards for good list behavior. Unfortunately, moderation is a good deal of work: a rent-a-moderator service might be a good business idea, locating the moderators in low-salary countries.
Some people may say that doing so would encourage users to get off the list and thus lose customers. But somebody who is getting tired of your emails is not likely to be much of a customer anyway. And the more you annoy that person, the less they will like you. Make it easy to get off, and maybe they will get back on some other time.
Most important, if you have any belief in the concept of permission marketing, you should recognize that you only have users' "permission" as long as they actively want to hear from you. Permission does not mean "I once conned the user into giving out his/her email address, so now I can do as I please."
My preferred unsubscription design is to create a dedicated email address for each user: if any email is sent to that address, the user is removed from the mailing list. This personalized unsubscribe email should be listed at the bottom of every single mailing list message.
When users unsubscribe, they should get an acknowledgement message that confirms that they have been removed from the list. This message should include the instructions for how to subscribe:
A few months ago, I changed mailing list provider for the Alertbox announcement list, so it serves as a good example of the impact of usability on subscription growth:
The old list server was particularly bad and I only used it because it was (a) free and (b) maintained by somebody else. It used the old-time batch processing command model where users had to type exact commands in the body of the email messages. Even worse, the confirmation messages for successful subscriptions were sent from the address alertbox-errors. Yikes: way to spook your users. And to activate their subscription, users had to reply to this message and edit the subject line to indicate their approval. Many failed to do so.
The Alertbox announcement list is now hosted at SparkLIST, an ASP that specializes in high-end mailing lists. I have been very happy with their service. The usability of the interface for list administrators (i.e., me) could be much better, but the UI for subscribers is good.
Response time requirements for email are quite different than those for Web pages. As mentioned many times, Web pages need to download in less than 10 seconds to satisfy basic human factors requirements (and 1 second for optimal usability). Email does not involve navigation and users don't sit and wait for individual messages. Thus, most email can easily take several minutes or as much as an hour to be delivered without any usability problems.
Delays of more than an hour may cause users to feel left out of a discussion group. Or they may feel that the news is old when it reaches them. The more a list depends on news, the faster it must be.
The one exception is email that is sent as the direct result of a user action. For example, confirmation messages should be sent with a delay of no more than one minute or users will start to worry that the system is down or that they have made an error.
The services currently provided by mailing lists should be shown in a communications control panel that would monitor all the things the user is interested in on the Internet. Areas that are hot or have news would be highlighted, and an agent in the user's computer would prioritize the possible sources of information and show the most important ones with the most real estate. For sure, the communications control panel would collapse threads of discussions into a single object and visualize its activity in a more useful manner than hundreds of scattered lines in your inbox.
Wilson's Web Marketing Info Center has a long list of additional readings about email newsletters, complete with links and summaries.
Copyright © 2000 by Jakob Nielsen. ISSN 1548-5552