How did that last outreach campaign or template perform? It’s easy to find out; just check response rates or count up all of the links you earned.
But smart marketers want to dig deeper: What sort of outreach is working best for the company, overall? What are the best practices when doing content outreach? How do I write a good subject line? These questions are much harder to answer, and most of us rely on our intuition and experience to do so. Today, I’ll show you how to instead answer with data.
Step 1) Brainstorm
Make a list of what you’d like to measure. Include the things that experts tell you are “best practices.” Include things you believe to be true or not to be true. Think about the questions that your boss and your colleagues always ask you. Include everything.
Here’s what my list might look like:
- Outreach tactic/type (e.g. content, product review, announcement)
- Subject line length
- Message length
- Use first name in greeting?
- Refer to their blog name or blog URL? In subject? In body?
- Use our brand name? In subject? In body?
- Use our URL? In subject? In body? In signature?
- Subject line is casual? Is vague? Is formal? Is data?
- CTA: Ask for link? Ask for share? Don’t ask for anything?
- Has a compliment? Talks about a recent post or article? Doesn’t talk about them much at all?
- Description of our brand/product/site or no description?
Step 2) Organize
Next, organize the brainstorm into an outline. Give each research question its own line. These will be your variables. I find it’s helpful to sort them into categories, too. I usually use message type, subject line, and body.
Also, make note of what type of answer each question will need. In my example above Brand Name in Subject is a Yes/No question; Outreach Type requires picking from a few categories, and Message Length is a count.
Step 3) Build Your Spreadsheet
Now you’re going to create a spreadsheet to track all of your variables. Your first four columns are where you’ll enter your template/campaign information. They are: Subject, Body, Sent, Response. (Add a fifth column for Linked if you have that data.)
After that, each variable gets its own column. I usually make a short variable name in Row 1 and then describe the variable in detail in Row 2.
It should look something like this:
Once you’ve got your template sorted, you can start entering campaign information. This is often the most tedious part of the process, so grab a coffee and find a good playlist to get you through it:
Step 4) Format the Data
Here’s where the fun begins. Take your spreadsheet and make it a table. (You can delete the row with the variable descriptions now.) In Excel, select Insert -> Table or use Ctrl + L.
Now, make it a pivot table:
Step 5) Start Analyzing
Place the variable you want to look at in the “Row Label” section, and add “Sent” and “Responded” as values. You can add a column to calculate response rate if it helps you mentally process the results (That formula is =responded/sent).
Step 6) Check for Statistical Significance
In the above example, it looks like emails that use a blog’s URL in the subject line perform much worse (11% response rate) than emails that don’t use it (56% response rate).
This is good information, but we need to figure out whether or not the difference between those two numbers is actually statistically significant. To do so, you can use chi-square calculators in Excel, or your can plug your data into this calculator by Rags Srinivasan.
In this case, I learn that, even though 56% vs. 11% seems like a huge finding, it’s not statistically significant. Therefore, I can’t really say whether or not it’s a good idea to use a blog’s URL in the subject line. This data set is telling me that it doesn’t make much of a difference either way.
Step 7) Take Lots of Notes
Keep a running list of what you’ve tested and what you’ve found. Note all the variables you looked at. Mark the ones that were statistically significant, and write out corresponding insights for each one.
Step 8) Share Insights
Share your findings with the rest of your team. Create a list of best practices that you can refer back to and/or use to train others. (Here’s mine.) Incorporate your findings into future outreach templates.
Notes & Conclusions
If you have a large enough dataset, you can pivot by person or by message type. This will help give you an idea of what types of response rates to expect when you’re planning future outreach campaigns. You may also find opportunities to recognize high performers or to train lower-performing folks.
This is very much an imperfect science, and it’s not meant to provide black-and-white results. (It’s research, not reporting.) Use it as an opportunity to figure out what’s working for your team and what isn’t. Hopefully, you’ll find some things that surprise you. (Let me know if you do!)