Marketing School: Beginner’s Guide To Doing The Splits
Did you like that title? Isn’t it search engine optimized? (It’s definitely optimized, but I’m not sure for what.) That is because I am an excellent and well-respected marketing professional. In other news, thank you for all of your emails asking if I’m dead. I’m not.
If you have a home business marketing consultant, they will often advise or supervise a split campaign — often called an A/B split.
This sounds scary. It’s not.
When I first heard this, I freely admit I freaked out. I will always associate A and B with algebra (X and Y I associate with chromosomes) and I don’t dig algebra. I am not saying this for comedic effect — I was borderline hysterical. Granted, I become borderline hysterical when I can’t calculate the tip at my friendly local tavern, but this was worse than usual. Anyway, moving on.
Splits are simple. Take your ad or your sales letter or your landing page or your spam or font or whatever, make another version that is very slightly different, and send them to two groups of equal size. Then measure the response. Easy peasy.
The reason for this is simple. When it comes to copywriting, either you’re doing it yourself — and let’s be honest, you probably don’t have a clue what you’re doing — or you’re paying someone else a hell of a lot of money to do it for you. In either case, figuring out what works as quickly and as accurately as possible is of the utmost importance. (According to MS Word, “upmost” and “utmost” are synonyms. Who knew? And I call myself a copywriter.)
Here’s what you need to know about splits, A/B, X/Y, and otherwise.
Make the difference between A and B very slight. If A and B couldn’t pick eachother out of a police line-up, you don’t have A and B. You have A and L. A and L is bad because if 90% of your sales come from L, you don’t know why. You can’t tell exactly what it was that made L better than A. (Actually, if 90% come from L, just go with L. It’s obviously fine.)
Make the difference between A and B at least marginally important. Do not change “very” to “really” in the third paragraph from the end. Any difference in effectiveness will be coincidental and not representative of a trend. If you’re new to splits, headlines are generally a good start.
Make everything else as equal as you can. If you’re doing direct mail, alternate houses. Don’t send A to a rich neighborhood and B to the projects. You need to keep everything consistent.
For God’s sake, measure it. Set up a system that can let you know which sales or leads are coming from where. This is a sadly and surprisingly easy mistake to make.
The best A/B split is one where you can hand both versions of your copy to an impartial observer, ask them which they like the best, and they don’t really have an answer. (Of course, if they don’t have an answer because both versions are shit, well, sorry dude.)
What kind of splits can you run? AdWords will monitor A/B split campaigns for you. You can have slightly different landing pages. You can have different colored fliers. Hell, you can even have different prices. Pretty much anything you want, you can split.
Some Common Splits:
Ever see a TV commercial run on a major network — one that you know isn’t a local feed — and the ad says to “Call Sue at …”? On the other commercial, they’ll say to call Jane, or Mary, or whatever. They want to see what ads are making people call. You call and say, “Yo, I wanna talk to Sue” and they know you were watching Dancing With The Stars at 9 on Wednesday in Duluth.
How about coupons? When you get a coupon in the mail and it has a numeric or alphanumeric code, it’s usually because they’re running a test. Some poor sap in the data entry department types those bad boys in and then The Powers That Be In Marketing determine what got the most asses in the seats.
Online sales and information products are notorious for splits because they’re nauseatingly easy to do, with the potential for astronomical profits. If I’m selling my ebook to you guys for $29, I can run an AdWords campaign to see if it will sell for $59 and keep tweaking until I hit the optimal price.
In other Search Engine Optimization news, I’m pleased to announce that no less than 25% of my search engine traffic typed in some combination of the words “Naomi” and “thong”. Because I’m cool like that.
Liked it? Click here to subscribe.
***
Want a small business marketing coach of your very own? Click here to get started.
Next Post: When You Feel Like A Raging Failure
Hey, often times people forget about these basic marketing ideas…I know I do. I wrote a similar post(more aimed at split testing ads) the other day but after reading yours I think I missed on some good points. Thanks for the info.
-Ryan
Ok, I am a dumbasss. This is very good but at the same we often fail to do it. Do I have to admit that last month I did this exact thing — and then failed to really measure the difference response rates? When I did it I even told myself to tabulate the responses, but then I just went “aw heck, I will eyeball it.” Dumbass. Luckily it wasn’t anything important, but it really would have been nice to know the results.
Michael, I think that officially puts you in the same boat as 99% of everyone else.
Which is why split testing is cool. Enough people don’t do it that it can give you a big ole unfair advantage if you do. Bing! You are a marketing genius.
@ Ryan - Hey, so you did! I’ve never seen your blog before, so thanks for the intro.
@ Michael - Ditto what Sonia said. It’s awful easy to forget that step.
@ Sonia - Bing! (I don’t really have anything intelligent to say right now so I’m going with “Bing!”)
totally misleading title. i thought this was an article about strippers.
This may be a dumb question, or very apparent to someone who knows a lot about marketing, but I’m intrigued by the idea of testing the waters to find the optimal price of what someone will pay for a particular product/service, say like you suggested using Google Adwords. But don’t the people who bought it for $50 get upset when they find it a day later on your site for $20. How would one get around that? Logistically, I mean….two separate sites?
Hi Janelle - Yes, two different sites. Two different landing pages that can’t be accessed from the other is a good idea as well. Also, when it comes to splits on information products, it only really works if you don’t have a high level of brand awareness in place. Freelance Switch couldn’t put Rockstar Freelancer at two separate price points because their readers would notice. They could, however, advertise it without the branding and with a slightly different title through AdWords to see if people would pay more. (Ideally this would take place before the media launch.) :-)
@Naomi -
What a coincidence! 25% of *my* traffic comes from “Dave Navarro” and “thong!” But I do NOT want to go there.
Split testing is invaluable. If it seems intimidating (like it did to me), you can go with an inexpensive web service like HyperTracker (like I did) that’s simple to set up and gives good, real-time stats. I pay $20 a month for it and it’s well worth being able to set up split tests with about 5 minutes of work. (If anyone else knows a good service that’s cheaper, let me know & I’ll jump like a bat out of hell).
@Michael/Sonia -
That’s the reason I use the web service - no need to eyeball or go thru web logs. It’s also opaque (as in it’s not possible for a user to tell how to get to any of the pages in a spilt test).
Cool, Dave, I will check that out.
Bing!
[...] I’m a Web Developer. I just happen to work for myself. I have found through experience that people look on me more kindly as a Web Developer than as a Freelance Web Developer. I tried both titles to see what worked best and you should too. It’s a marketing trick called an A/B Split. Naomi from IttyBiz has humanised the idea by cleverly dubbing it “doing the splits”. [...]
Another great tip from Naomi, who I am glad is still alive.
Great post. I do something like this when I am marketing a new site. I take two or three different pieces of copy and place them in forums and other places to see what gets the best response. When I see what ads bring in the most, I do more of a mass marketing campaign with the winner and the tactic has not failed me yet.
This is an informational post. Having a great copywriter put together a couple different pieces together is great when you want to be positioned to different markets and see what message resonates the best.
[...] Marketing School: Beginner’s Guide To Doing The Splits - IttyBiz [...]
[...] overlook the occasional use of a semi-colon. Because of Naomi, I (somewhat) understand things like A/B splits, press releases, target demographics, and unique sales propositions (USP’s). Just remember, he or [...]
[...] overlook the occasional use of a semi-colon. Because of Naomi, I (somewhat) understand things like A/B splits, press releases, target demographics, and unique sales propositions (USP’s). Just remember, he or [...]