Give your next digital campaign laser death beams
While continual optimisation is admirable, Ryan Sproull argues marketers and agencies need to look to do something radical to really evolve.
I’m a big fan of X-Men.Yes, this is going to be one of those digital-strategy articles that start out talking about comic books. You know the ones.
Sproull
The idea in X-Men is that “mutants” are people who were born with a particular gene, called the X-gene, which gives them superpowers. Pretty much any superpower you can imagine, as long as you can come up with a pun for a name to go with it.
What I like about this idea is that just the right kind of change creates radically new possibilities. Animals usually evolve in small increments – a little taller, slightly better eyesight– but with the right kind of change, you don’t just get small variations on the previous theme.
You get folks with laser death beams blazing from their eyes.
Evolution and Optimisation
In the business world, evolution is the perfect analogy for the optimisation of our use of digital assets and channels. The most common form of optimisation is A/B testing. That’s where you try two variants of something (e.g., email subject lines) and compare their performance. The winner is whichever performs better in terms of whatever metrics you’re hoping to achieve (e.g., email open rates). And then you keep repeating that process with variations on the winning variant.
Optimisation is also called “continual improvement”, and that’s exactly what you get: continual, gradual improvement that’s responsive to the changing environment of your audience’s desires and expectations.
But what you won’t get is a guy with laser death beams blazing from his eyes.
That’s the problem with microevolution. You get improvements on existing themes, but no new themes. You get better and better hearing, but no mind reading. You get faster and faster runners, but no teleporters. You get higher and higher jumpers, but no flyers.
Happy Accidents
A while back, I met some people who ran an e-commerce website and were avid optimisers. They constantly used A/B testing to refine their site’s homepage – trying out differences in layouts, calls to action, imagery, navigation labelling.
They would select for whichever variant gave them the highest rate of conversion to purchase, and build from that. Small gains of 3 per cent here, 4 per cent there.
One Monday, they came in to the office to discover that over the weekend their conversion rate had leapt by 60 per cent. It turned out that their A/B testing platform had crashed, reverting the homepage to a basic default that had no menu at all.
Visitors were forced to use the search bar, and for whatever reason, searchers were more likely to convert to purchase than menu-users.They had learned something incredibly valuable, by mistake. And it had to be a mistake, because if everything had gone according to plan, they would never have made the discovery. That’s the thing about sensible A/B testing:
The same optimisation approach that makes progress inevitable makes brilliance impossible.
A/B/X Testing
I’m not suggesting you stop optimising. And I’m also not suggesting you sit around hoping for a paradigm-shifting calamity to give your business superpowers. What I propose is an approach of gradual improvement that consciously keeps the door open to the possibility of radical improvement. I call this approach “A/B/X testing”.
The X stands for X-gene. Or the Roman numeral for 10. Because I am proposing that alongside your usual optimisation regime, you set aside 10 per cent for trying completely new things. 10 per cent of what? 10 per cent of your resources: your budget, your time and your audience.
Expose one-tenth of your audience to something completely different from your usual activity, something you produced with one-tenth of your usual resources.
Your X-variants should be thoughtful, directed, but experimental. They should give creative people the permission and space they need to play. They should be designed to test hypotheses, question received truths and assumptions, and they should be designed to provide learnings even when they almost inevitably fail.
And fail they will.
Until, of course, they don’t.
And that’s the difference between focusing on tactical metrics and focusing on strategic business objectives. By the short-term standards of success of a campaign or landing page or display ad, it can seem like you can’t afford to take risks with radically new ideas. But by the long-term standards of success of your business, you can’t afford not to.
So you want to be a mutant…(I’m assuming you already have an optimisation regime in place. If you don’t, drop a line to your friendly neighbourhood digital strategist, because you almost certainly should.)
If you’re a client, talk to your digital agency about the idea. There’s a good chance they’ll be keen on it – all of the best agency folk live for the opportunity to think outside the box on a project.
If you’re an agency, you’ll probably want to pilot things with a long-time client who trusts you enough to commit budget to something that will almost certainly reap no short-term gains.
Finally, look for something small to start with – an upcoming campaign’s landing page, a business-as-usual SEM spend, some EDM activity. Get a feel for what processes and resources you need in place to support it. Big discoveries can come from small experiments.
Ryan Sproull is a digital strategist at Zuni
WHAT
User ID not verified.
Thanks for the article.
Agree – great to have the wild card in the mix. This is what I love about optimisation – a safe way to play the wild card – use it only with a subset of traffic to begin with, and see how it does.
More incremental A/B is still great. Knowing what small changes make a big change is still really valuable. But, it is nice to give the creatives some love and test out the big hairy idea and see how it plays.
The problem with introducing the hairy wild card – It is hard to figure out ‘what’ about that version is what people to respond to, and still need work to sort out ‘why’ it worked.
User ID not verified.
This article really confuses me. I’m with Wendy…
User ID not verified.
Completely agree, Scott. We only need to keep the door open to wild cards; definitely don’t want to be pursuing them exclusively.
Regarding the danger of taking the wrong lesson from a happy accident, excellent point. I guess the best answer is more testing – test hypotheses for why something worked before declaring it a magical learning.
The last thing you need is to think a landing page worked really well because people respond well to red and green together, only to discover that actually it’s just that colour-blind people happen to adore your product for some reason.
User ID not verified.
Visionary!
User ID not verified.
I kinda figured some agencies were already tinkering in this area? In another life, before I saw the light and mutated into a digital copywriter, I was an above the line guy. I remember the creatives at BAM saying they always presented three ideas to the client: two being safe and the third way out there, with the hope of doing great business and maybe bagging an award or two. It was all about where they could take the opportunity. Since then I’ve worked at stacks of agencies with a similar attitude. Again, it’s searching for what one can do with opportunity. Done right, it’ll be the client that thinks you’re the super hero!
User ID not verified.
Nic,
Yep, plenty of agencies do the two-safe/one-brave pitches, or some other variation on including a bit of madness in the mix.
There are some important differences between that and A/B/X testing, though.
Firstly, the tactical objective of each. In your scenario, the tactical goal is to produce something good, perhaps great, and maybe win an award. In A/B/X testing, the tactical goal is to learn about the audience, the market, the environment. In both, the strategic objective (bigger picture, beyond the success or failure of a single project/campaign/action) is the business doing well.
In your scenario, most of the learning is about the client – what they like, how risk-averse they are, etc. If nine times out of ten the client picks a safe option, nine times out of ten you learn nothing about how well or badly new things work in the field.
Secondly, in your scenario, all of the eggs are in one basket. If the client trusts you with a risky idea and it falls flat, that’s a big deal and you’ve only learned a few things. In A/B/X testing, you’re putting a tenth of your eggs in a basket that’s ten times smaller and you learn 10 times as often. Or something.
The two aren’t mutually exclusive. A/B/X testing is about formalising a regular habit of low-volume high-risk actions for the purpose of learning (while continuing mostly with the safe testing, as Scott mentioned above), and that can be going on during and within campaigns/projects whether they’re considered safe or adventurous overall.
User ID not verified.
Interesting article Scott that is clearly written and easy to understand.
Whilst I agree with your overall point and conclusions, you could also interpret the results from the example you gave in a different way.
f the e-commerce site had more insights into the user experience, they would have a better handle of what areas of the site are more likely to contribute to the overall conversion rate. If they then did A/B testing on those areas first, I would argue that they would have got more significant improvements quicker as a result. Otherwise, you need to do A/B testing on everything with the occasional x testing.
There are many other types of testing that they could have done both qualitative and quantitative to give them this insight.
Whilst I agree with X testing, I don’t think it should substitute for insights.
User ID not verified.
Trying and testing not only new creatives and buy types; but entirely new challenges is, in my opinion, the key to successful optimisation.
The truth is; we are no longer living in a single channel, or even ‘multi-channel’ digital age any more. Where marketers should be moving towards is a true cross channel approach, because this is the age we are now living in.
Where in times gone by, running multiple digital channels was the norm; the insights and audience habits learned in activities in each silo’d channel were only really played out in future optimisations to each of those channels.
However, running digital channels not in isolation; but truly fusing the data and audience insights from one channel to another; this is the way forward and this is what I’d call true cross channel.
I’m also a huge advocate of this concept to ‘allow 10% of your buy budgets for investing in new things’; because it ensures you are always adding new channels into the mix and learning from these trials.
User ID not verified.