I want to thank all the agencies out there who get Portent new clients. Thank you. When you promise a client you’ll send them ‘monthly reports’, and instead e-mail them something like this:
You couldn’t be helping me more.
Bless you. You’re totally disappointing your client, because they understand that a report is not the same thing as analysis. And, they understand that they’re paying for analysis.
If you don’t like sending me business, I suggest you learn to write reports:
Crunch the numbers and do your analysis first.
I used to think this was obvious: Don’t write your report until you know what you’re going to say. Alas, I’ve been proven wrong several times.
I usually put together a spreadsheet, first. It includes all of the raw metrics plus a brief analysis.
Then I turn that into a rough outline. I don’t usually do outlines, but for a report I find them helpful.
Duh. But don’t just cut-and-paste nasty-looking charts. Don’t even cut-and-paste beautiful charts. Create and format the data yourself. That way, you can put those diverse data sources together. You can also illustrate trends, annotate and clarify the information.
Tell the client why the data’s important. This is where most report writers fall apart. They’ll say stuff like:
“Traffic went up.”
“Sales went down.”
“You got more traffic from the word ‘fudge’.”
Great. But why does that matter???!!!!! If I wanted to know that ‘traffic went up’, I could look in Google Analytics.
I want to know why traffic went up, and what impact it had on my web business as a whole.
Provide analysis. Analysis is your first opportunity to lend something to the discussion.
Provide a next action.
The data and analysis is useless if it results in nothing. You want to show your client the next step.
Wrap up each bit of analysis with a ‘todo’—some recommended next step based on your data and analysis.
Also, don’t wait for the end of the document to make your recommendations. Make each recommendation after each bit of analysis. That puts the recommendation in context.
Correlate diverse data sources.
Pull different metrics together and see if there’s insight. Show the client how no, links didn’t go up with their last link bait piece, but yes, social media mentions went up, and traffic from organic search went up, too.
That kind of correlation is what separates you from Google Analytics and Chim Chim the reporting Chimp. You can actually correlate data.
Compare like with like.
If you show multiple charts together as a comparative tool, use the same axes and ranges for each one. Or make it very clear why you’re not. Otherwise, your data could be very deceptive.
No 3d charts. No flowing descriptive prose. Get to the point, fast. This isn’t a 7th-grade creative writing assignment, it’s a report.
Say you’ve just reported bad news, like:
“Sales dropped 15% after we shut down the PPC campaign.”
Following it with something like:
“But SEO is up nicely, and I’m looking forward to a good 4th quarter.”
Is like saying:
“I suck, have failed you utterly, and have no future. Abuse me as you like.”
Just give the message and move on. If there’s a solution to the problem, suggest it. If it was a mistake, admit it. But move on.
Use a real word processor.
I once had a client who complained I wasn’t sending him any data. Thing is, I was sending him a report every week. But each report was just an e-mail. He’d read the first 2 sentences and delete it. Make your report a Word doc and suddenly everyone reads it. I don’t make the rules. I just explain ’em.
Use a real word processor, the right way.
Use the damned template! If there’s a preset style for ‘heading2’, use that for your second-level heading. If there’s a preset style for a table, use that, too.
Doing that will let you generate a table of contents, table of tables, list of images and cross references with a click. Failing to do that will make your boss want to maim you.
Write for 9th grade, at most.
The reader isn’t going to have a lot of time. Don’t use ‘trended upwards’ if ‘increased’ will work. Some other favorites:
“Underperformed” could be “lost money”
“Bad strategy” could be “mistake”
“Budgetary limitations” could be “budget”
I’m starting to grit my teeth, so I’ll stop there.
Break up the text.
A lot. A great report will naturally have a mix of charts, text and tabular data on nearly every page, so the reader can get a break from reading paragraph prose. Even if that doesn’t happen naturally, break up the text the same way you break up text on a web page.
Bring it back to business goals.
Always tie your analysis and recommendations back to business goals. Remember, you live and breathe the web site every day. The CEO probably doesn’t, unless you’re a web company. And even then, she’s not in the details the way you are.
If you can bring each finding back to overall business goals, you put it all in a context that makes sense to everyone at the company. And, you’re doing what a good marketer should – giving communications direction to every level of the company.
Provide a high-level preview at the start.
Lead off your report with a quick summary of the critical metrics, and what they mean. That lets the reader get an ‘at-a-glance’ look at how things are going.
Brag a little
It’s perfectly OK to brag a little. Be sure to point out what you and your team did that helped. Make the connection between your work and the results 100% clear.
But actually, please ignore all the previous tips. Your crappy reports, oh great crappy report writers of the world, generate at least 50% of my leads. Carry on.
- My director of search, Elizabeth Marsten, has written a small business PPC ebook series. She is the queen of PPC – you want to get this book. She’s also a gifted Kendo practitioner, and I don’t want to be beaten senseless.
- Great ideas from Guy Kawasaki’s Enchantment
- Fluffy rainbow bunnies devour marketing agencies
- The content revolution is here, and it’s full of crap
- Log files bring sexy back
- 25 tips for building your own agency
- My SEO copywriting e-book – $7