The Future of Advertising Is Anticipatory
I wrote this piece two years ago for AdAge, it was one of their most popular posts ever.
Now at a time where AI is on everyone's lips, whether it's Google Assistant, or suggested content from Netflix or what Amazon will soon become, this seems like a good time to revisit it.
Retargeting Is Flawed; the Future Is Pretargeting
The Future Lies in Targeting Based on What We're About To Do, Not What We've Just Done
There is no time in my life I am less likely to buy some white pants, a toaster or a ?ight to Los Angeles than after I've just bought these items, yet that's precisely the time I see ads for these products or services.
These ghostly images stalk our internet journeys like shadows. While ineffective, these ads come to us by some of the most advanced technology there is. By some measures, they are the most appropriate ads to serve us; they can be the most noticeable, but they are also the most pointless. It's like an inaccurate watch -- knowing the engineering inside is impeccable doesn't help if it can't reliably tell the time.
Retargeting is based on the past. The clumsiness of retargeting is well-documented and we're assured that teams of scientists are working away to improve the algorithms -- these are the new media buyers, planners and creatives of the performance marketing age.
I have no doubt that retargeting will get better. When more and more behavioral data is overlaid with checkout data, credit card data and recommendation engines, we will soon see a new era of personalized advertising. We'll be shown ads for big-ticket items at precisely the right time, after we've been thinking about them for a predetermined period. We'll be shown ads for items that work perfectly with our new white trousers. Technology is moving so fast that soon both the ad placement and the advertisement itself will become automated.
We'll see fully rendered completely personalized video ads based on real-time pricing, real-time availability, the weather and thousands of other data points. The art director and copywriter team of the future is the algorithm and processor.
Nonetheless, the main problem with retargeting remains that it's based on the past. It's about reaching people who have bought items, who've accepted they need something, and who have started to act.
From the past to the present
Search focuses on the present. The value of search is predicated on its future vision; search is the only opportunity in advertising to address people at their moment of need, before they take any other action to quench it.
We search at the start of the purchase funnel. Search will for this reason alone forever retain incredible value in the advertising mix -- the single best way to convert known interests into action. But search requires us to know that we need something. It takes effort; sometimes we need help and we don't know it or want it. Predictive ads are about the future. Search will soon be trumped by something mind-blowing -- what about converting needs we don't yet know we have?
Predictive advertising: From the present to the future
The digital ecosystem is an untapped mine of data for advertising, and our changing behaviors are making it even more so. The more we like posts on Facebook, accept location services, connect our address books, use our digital calendars and the more we share this information across devices, the more accurate the pictures of our lives become. When companies with incredible processing power overlay what we do, where we go and how long it takes -- both as individuals and collectively -- a rich new picture of what we do emerges.
The world of predictive computing is upon us. We see a few early entrants -- we have Google Now predicting our journey to work and telling us to set off early if traf?c blocks our normal route. Microsoft Cortana may see our upcoming ?ight to Sydney and offer relevant information like exchange rates and weather forecasts. And it's this value that changes everything.
The current privacy debate is focused on privacy being lost -- the debate needs to move to privacy being exchanged. Facebook's recent decision to announce its new privacy terms but also to explain how consumer information would be used to make better advertising seems rather sensible. It seems like a value exchange. Aided by a degree of opting in, we will likely soon see a world of advertising experiences on another level -- the world of ads for products at the right time, the right place, with the right message all based on predictions about your needs at that time. It will be the coffee shop suggested when you have time and it's about to rain; the Taxi2 ridesharing app telling you someone else wants to go the same way as it thinks you do; the car garage that will repair the problem only your car knows it has; the train times when snow blocks your route to work; the meal special offer when your friends are close and your calendar is free.
Future advertising will soon become a valuable offering of useful information at the right time and at the right place and in the right way. This contextually precise assistance, aided by a landscape of new connected wearable devices to provide better inputs and ways to convey information continuously, changes everything.
Predictive advertising should be an incredible emerging world for brands to exploit. The technology is here, but what's stopping it is the huge cultural shift that needs to happen that will allow more comfort with sharing information. Predictive advertising comes from trust, and when we accept that privacy is a lost battle and decide to embrace this, then we can get something in return.
AI-First Organic Search Specialist
8 年Very interesting piece, Tom. I couldn't agree more that the privacy/data conversation needs to switch to the topic of value exchange. Consumers have shown that they want the benefits of being 'connected'. The only thing holding them back is that niggling question in the back of their minds: "what am I giving up in return?". The best place to start would be allowing consumers more control over the outcome of sharing their data. I've always thought that the functionality that will cement the relationship between predictive advertising and consumers will be the ease of two-way interaction with it... the ability to 'mute' or opt-out of certain topics at certain times. Predictive might not always get it right. If I buy a ticket to Sydney, then understandably I might start to see ads for currency exchange or tickets for the Opera House. But I might just have a bunch of Australian dollars lying around from a previous visit and be planning a cheap, no-frills trip on a shoe-string budget. It's easy to imagine a scenario where algos would not be able to deduce those things. But, if I could hit a button on those ads to say "don't need this" or "not what I'm looking for" (perhaps even "not in my price bracket") then I could still feel in control. Future ad space would then be freed up for more and more relevant comms. Conversely, a button to ask for "more like this" would also be a win-win for both advertiser and consumer... where I further prequalify myself during my predictive ads journey. While the passive and social inputs for predictive algos are enticing to the developers of the tech, I think that consumers would respond much better if a certain level of manual control still remained. Predictive will likely live or die by the balance struck between those two types of input. In the new age of relevance, we'll be fighting a brand new conception of 'spam' where being only just off the mark by a whisker may be enough to evoke negative feelings from an ad.
Founder///Future-Hacker
8 年Until we understand that privacy is as fundamental to the human condition as clothing we will not see that the real opportunity lies is in creating a value exchange that puts self sovereignty and agency in the hands of the consumer at the center of the brand engagement ecosystem as an active agent as opposed to as a passive target. The future of advertising will be powered by intent-casting, actively cast by people to the brands that they trust via AI agents that work on their behalf to ensure that information and intentions shared and the value extracted from them serve their owner's needs first and foremost. Anticipatry advertising will be inept, ineffective and unwelcome until it embraces intention as a direct active input from consumers or their agents and creates a value exchange that fairly rewards people for sharing this high value elixir like data as opposed to trying to extract it from increasingly invasive, extreme and surreptitious means.
Digital Marketing Director
8 年Targeting should instead of being based off of the past, be based around products/services that immediately precede or proceed the product/service you just bought. No one wants to see ads for things they've just bought, but for products/services that compliment what they just bought, now that's valuable.
Corporate Sales and PR
8 年Ah very interesting. Predictive sounds better to me than right time or retargetting. Amazon does predictive already as does IMDB as they group us on the basis of liking similar content. Google also can go predictive with their affinity groups. I think the safest way in a return way is to create context. So even if I don't want that airline holiday today, when I do want it, your ad is in my head at a subconscious level
AI & Emerging Tech Strategist | AI Governance Professional | Fractional COO/CCO
8 年Good article and insights. I am working to shift the way we talk about data and sharing access to our data - more than talk about Privacy. We have data. We provide access to that data when we have 3 things: Trust, Value and Control. Is there sufficient value being provided for access to our data? Do we trust the entity or individual to whom we are providing access? And do we have control over where the data goes and how it is used? An example is my health data. I am happy to provide this data to my doctor because I trust him, he will help me to stay healthy by using the data (value) and I know it will stay within his office. However, that very same data I do NOT wish to provide access to an insurance company. I do NOT trust them, they could actually create negative value (raise my premium) and I do not feel as though I have control over where and/or how the data will be used. My point is that we need to have more than just a value exchange. We also need to be open about the where and how the data will be used, give consumers the ability to control this and to ensure that we act in ways that build trust. It's much more complicated than just "privacy"