Lean Validation via Email

Engineering time is valuable. Don’t waste it on something you aren’t sure has value or that you aren’t sure what should be included. This is especially true with emails. Even with libraries like blueprints from Mailchimp, coding up HTML emails is the least favorite thing most software developers do. Therefore, make sure that the emails you are asking them to create are both of value to the customer / team and fully baked ideas. I’m not suggesting that they be perfect and never need iteration, but don’t think that changing the copy on them every week or two is in your best interest.

Step 1: Understand the problem

Why are you even sending an email? Is there on an on screen option that could suffice? Even if the conclusion is still an email, understand what user problem this is solving. Ask questions like:

  • Who is the audience for this email? Are there different groups?

  • What do they want to know?

  • What do they need to know that they wouldn’t think to ask?

  • What do they already know? (what stage are they as a customer?)

  • When do they need the information?

  • What voice do we normally communicate to these users with?

It’s ok if you don’t know all of the answer to these question and you make some assumptions. As long as you call that out and decide when and how you will be validating those hypothesis.

Step 2: Create a single template

Don’t try to solve all the problems identified at once. Create a basic email template the covers what you can. Iterate on it… the subject line, the copy, the jargon, the images, the time of day of sending, who it gets sent to if there are multiple users within a group…

Step 3: Create a second template

Iterate on the first one and you’ll probably realize that there are separate user groups who have slightly different needs, create a second template to help with those problems. Create reports and decision trees that help you decide who will get which template. This will establish the business logic the engineers need to know. If you can’t create a simple decision tree, it will probably be hard for the engineers to code the logic as well, so simplify where you can. Iterate on the copy as well as the imagery you’re using based on the replies you get from users.

Repeat and expand your template library until you have the vast majority of use cases covered. If you have more than 3 emails, see where you can find common ground and consolidate. Again, don’t strive for perfection, it’s ok to have a “Here’s who to contact with additional questions” sort of ending.

Step 4: Automate the email

Use the business logic and the copy and images to generate the actual emails to be sent by the product.

Step 5: Continue to monitor open rates and ITERATE

Your job is never done… stale emails suck… schedule a time to review.