- ARPDAUPosted 5 years ago
- What’s an impressive conversion rate? And other stats updatesPosted 5 years ago
- Your quick guide to metricsPosted 5 years ago
[Gamesbriefers] Are sexist ad campaigns an inevitable result of A/B testing?
Quick note: the image to the right is a genderswapped parody produced by staff at free-to-play game studio Meteor. The story behind it is amazing. You should go read it at the Hawkeye Initiative.
I wrote a post about Plarium’s marketing techniques for Stormfall, where scantily-clad women are used to market what is a run-of-the-mill midcore RTS.
I suspect that A/B tests have shown that a headline that implies exclusivity and an image that implies titillation leads to improved click-through rates, and that Plarium thinks that this is the optimal customer acquisition strategy. Are they right? Is Facebook and mobile marketing on an inevitable trend towards lowest-common-denominator (and often sexist) acquisition campaigns?
It’s certainly nothing new that “sex sells”. I think it just comes into extremely sharp focus in Facebook ads because of, as Nicholas noted, the extremely fine granularity of data that would be much harder to collect through traditional media (television, radio, billboards, etc. are all very difficult to track clicks/actions from). We certainly see companies who embrace that philosophy in traditional media (Axe body spray, Herbal Essences shampoo in the 90s, and plenty of others), but most companies choose to be more responsible or conservative or focused on their brand and image and instead produce ads that they feel are appropriate and representative of their company and product.
Perhaps Facebook, Google, and other targeted ad networks exaggerate this through individualized ad displays, mostly automated (and thus less curated) content, and and a sense of reduced accountability for companies who don’t feel like they are building a brand but instead just getting clicks. While immediate fault lies with the advertisers (and potentially with those clicking on them), the platforms also need to decide what they want the experience to be for their users. I think it will probably require the platforms to implement some more content control to really filter out these types of ads. Clearly they already have rules against nudity in ads and if they see fit they can either provide tighter regulations or continue to allow users to filter and flag ads they find objectionable to customize their experiences.
The issue here isn’t A/B testing methodology as much as the variables that are being altered.
I’m not convinced that you will always arrive at an image of a scantily clad woman if you literally try images of everything and select and refine for click-through only.
At some point in this process a nefarious person needs to suggest ‘can we add one variant with tits?’
My money would be on their lacking not only a sophisticated marketing analytics infrastructure but also the in-house domain expertise to properly analyze such an experiment. I can tell you without looking at any numbers that a half-naked woman on an ad will drive CTR higher. But CTR, in freemium, is the wrong thing to optimize for. Those new users are expecting half-naked women in the game now; are there any? Won’t they churn away once they realize that? And if they do, wasn’t the money spent enticing them to download a game for which they have unrealistic expectations wasted?
At the end of the day, data tells us what we actually do, rather than what we wish we did. I’d be far more tempted to suggest that the correct conclusion is that enough people are susceptible to this kind of marketing for it to work, rather than blaming those who use it. The flaw is in the unwillingness to admit that, rather than in using such techniques. We may not want these techniques to work, but that doesn’t mean they don’t.
Well, that and the fact that the game features no semi-clad women and thus, as has been pointed out, will result in lots of churn. That said, I’d love to see how long people play before they figure that out.
As an aside, I am constantly (un)surprised by how few games are genuinely and deliberately titillating, arousing or sexual.
I see a lot of ads for games on Facebook these days featuring a busty lady, a promise of fantasy adventure and that the game is for 18+ only. It doesn’t really matter what the game is.
They remind me of classified ads in the back of subculture magazines (the ones that advertise “massage parlours” and such) and the ads I used to see in the back of comics of my youth (advertising many a you-could-have-biceps-too product).
What all share in common is that they have to tell a story in a 1-inch postage stamp format, and that tends to mean they have to be immediate, shocking and oriented toward lizard brain desires like sex or pride. They can’t be nuanced, complicated or artful because the eye just won’t see them long enough to take them in.
So I guess what I’m saying is that I’m not surprised that those kinds of ads are most effective in that format Nicholas. But I don’t think sex itself is the only effective way to make them.
There is a big difference between Facebook advertising and classifieds, which is superior targeting. In theory it should be more powerful and deliver a better audience over the long term for well-loved game genres, but my impression is that many studios don’t spend a lot of time trying to master it. It’s easier to just slap a porny ad together than to figure out whether Pathfinder fans might like your game.
Long term though, who’s likely to be the better customer?
I agree that the format is constrained and that limits nuance but as has been said the ’scandal’ is that the marketing team is lazy and badly incentivised. Any A/B testing going on here can’t realistically be measuring lifetime value.
This is the curse of our obsession with derivative testing and performance marketing where you don’t know what question you should actually asking. I’m a huge fan of testing everything and getting as much data as possible, but sometime there are secondary effects which can’t be measured directly. Testing is has limits, especially A/B testing. It can’t make you a better creative (or better person!) it can just tell you which of your bad designs is least worst.
There are smarter ways to use this 1” space than just to stick a pair of tits in it and a catchy call to action… but short term I can see why people fall into this trap. They want as many eyeballs as possible and by the law of averages they might just convert some of the people who clicked on it- if those users wanted real porn I suspect they wouldn’t be looking at it through a Facebook ad.
We are in such a luxury position in games to be able to fully leverage the online medium; any internet or mobile app can take you to directly consume our products. Marketing guys with physical goods would love that opportunity. Because its harder for them they have to work harder. So they find ways to create integrated campaigns which use multiple media/PR to build a lasting message, to build trusted brands. Then when they use a small 1” postage stamp it suddenly becomes extremely effective. You don’t have to use sex to sell a trusted brand… the brand sells itself. Isn’t it time we grew up as an industry and did marketing properly?
Good and crucial points made about correctly measuring CTR and LTV.
But regarding LTV and “how few games are genuinely and deliberately titillating, arousing or sexual”, look no further than any Japanese card game released in the past 18 months. When I was working in Japan it was precisely this sort of A/B testing that caused us to add juicy anime T&A to every ad banner/splash screen/app icon AND to integrate titillation into the core evolution mechanics, regardless of how artificially tacked on this was.
My personal favourite: Slime Elena
Of course there will be fall-out, in terms of player churn, ethics, reputation. And don’t forget the impact on your talent acquisition and retention – I am a female engineering product manager and I am obviously no longer working for a Japanese games company.
In principle this issue goes well beyond game design – if the sole and primary goal of your app or business is monetization, there will always be quicker ways to make money.
Personally, even if T&A to that degree was demonstrated to bring increased numbers of quality, spending, retaining players to the game, I still wouldn’t do it. I like having some pride in what I do.
(In this sense I am much like a cancer researcher, or a humanitarian aid worker. I think making games about pretend racing cars is right up there on the pride-o-meter.)
Slightly on a tangent, but relevant (I think). When I worked for a gadget magazine, if we ran focus groups, most men said ‘Oh, having models on the cover with not much on is so tacky, it demeans women, what do you think I am, some kind of SEXIST APE?’.
And then every time they did A/B testing on covers – e.g. putting just a gadget and no model on in newsagents in one city / district – those ones sold *awfully*.
This may have been a bit about some newsagents racking them with hi-fi magazines rather than with the lad mags, but it wasn’t the whole story.
Which in a roundabout way, maybe hints that a lot of men click on bannerboobs like crazy, even if they then are normal ‘oh I’m here for the game, that marketing didn’t affect me’ players thereafter.
Or not. How did it turn out for Evony in the long term – is it still popular?