Customer meme personas

Imagine you’re advertising to wealthy British men aged 65+, married twice, who own sports cars and dogs, live in castles and enjoy holidaying in the alps. Armed with this niche user persona you can more easily imagine what ad creative is likely to resonate with your target customer. Grant this profile a fictional name – let’s call him ‘Alan’ – and we can better collaborate with your team by asking “what would Alan do?” and “how would Alan respond to this?”. If your demographic and behavioural targeting is accurate, you should expect anyone in your target audience to respond similarly to your ads. You may be surprised to learn the Prince of Darkness, Ozzy Osbourne and the former Prince of Wales (now King), Charles would both fall into this audience!

Personas can be a powerful and practical tool for synthesising user research into relatable archetypes. What most people get wrong is an over reliance on demographics. Marketers like to assert claims about Millennials in presentations, but there are 72 million Americans that were born between 1980 and 1996 and their membership in this group predicts no more than their horoscope. In studies Millennials, Gen X, Gen Z and Boomers have less group cohesion (a measure of like-mindedness) than ‘people who floss’, ‘daily nut eaters’ or ‘IT professionals’. Charles and Ozzy may share demographics, but it’s their ‘memes’ where they differ. Knowing Ozzy is a fan of heavy metal or that Charles loves fox hunting is more useful in creating an ad that works for them. Demographics have their place, but it’s memes that drive behavior.

For segmentation to be worthwhile, it needs to drive a significant difference in behavior, because there are costs associated with segmentation. First of all creative ideation and production costs. Personalization of ad creative to tens, hundreds, even thousands of micro-segments, as became fashionable in the early days of digital advertising, can cost tens, hundreds, even thousands of times more than creating one ad for all. Typically this means spreading yourself thin across the creation of lots of ads, which decreases quality. Custom development of large scale personalization software is possible, but radically increases the complexity of managing your ad campaigns. Often you’ll find the best ad for one audience segment is the best ad for all. To paraphrase Anna Karenina, “The best ads perform in all audiences alike; the worst ads perform badly in each audience its own way”. If you’re running the same ad across multiple segments, there’s no real reason to keep them segmented.

Any act of segmentation is a declaration that you know the customer better than the platform in which you’re targeting: which in the modern age is rarely true. Facebook and Google have more information about their users, and more capacity to process it, with advanced machine learning algorithms to predict who is more likely to buy. Every time you split an audience in two, you’re halving the amount of data the algorithms have on each campaign. Typically at least 30 conversions in the last 15 days is needed per segment for ad platforms to be comfortable handling optimization, and splitting the audience too many times can quickly take you below that threshold. It also leads to a kind of dead weight loss: Audience A might have cheaper users to convert than Audience B, but if the budget has to be manually reallocated it can’t happen fast enough to capitalize on that opportunity, CPMs are inflated, conversion suffers, and budget is inevitably wasted.

About the Book 📕

← Back to Essays 📝