Good heavens – it has been a while.
For those who care, the preceding two blog posts (1, 2) provide somewhat of an update as to what the hell I’ve been doing for the past 3 years. So basically, if you’re my Mom, or the trying-too-hard insurance salesman who keeps sending me LinkedIn messages, see you back here in a few minutes.
For everyone else, let’s just dive straight in…
What follows is my “prediction” for what the digital traffic landscape will become over the next 2-3 years:
I say prediction in quotations, because this landscape already exists and is (by far) the fastest-growing segment of digital ad spend and marketing tech.
So it’s more a prediction as to how quickly it will hit critical mass, more than anything else. And anyone running paid traffic is probably already dipping their toes in the “future”, so to speak.
Anyway. Key up your imaginary drumrolls, because here comes the big reveal…
The Future of Traffic Generation is – Without Question – Intelligent Audience Modelling.
Now, let’s explore what that means exactly, and why you should care…
If you’re a clever little marketer, probably the first thing that popped into your mind just now was Facebook’s Lookalike Modelling capability. And you might be thinking to yourself that the author (yours truly) has truly gone off the deep end of irrelevance if he thinks this is somehow new… let alone “the future”.
What’s he gonna reveal next – posting on Squidoo?
Well, at this stage I may be the official hermit of internet marketing (fair enough), but the argument of today’s blog post is actually NOT that Facebook’s Lookalike Audience modelling is the penultimate apex of traffic generation.
Instead… my argument is that it’s just the primordial beginning.
And I believe this is strengthened by the fact that even in its early stages, Facebook’s ability to intelligently learn who its users are is what’s allowed them to successfully become the first actual contender to Google’s lofty throne – and in fact already unseating them in the battle for Mobile Display.
The key takeaway from that article being:
“…August 2012 was also the month Facebook made an inspired move with the introduction of Custom Audiences, anonymously matching advertisers’ first-party data like phone numbers and emails to users’ Facebook accounts for ad targeting. It would be hard to [overstate] the impact of Custom Audiences on digital advertising.
“Facebook quickly expanded Custom Audiences’ capabilities in 2013, first with the ability to expand targeting to Lookalike audiences, and then with the addition of consumer data from third-party vendors such as Datalogix, Epsilon, Axciom and BlueKai. Google only recently countered Custom Audiences and Lookalike Audiences with its version, called Customer Match, in the fall of 2015.”
In other words: It wasn’t the size of Facebook’s audience that gave it the edge on Google… it’s what they could do with it.
Up until releasing Customer Match, Google only offered targeted traffic based on someone’s current session (what they searched, where they were browsing). Facebook offers targeted traffic based on who they are.
There’s a big difference – and folks, this is the new land race. But it’s not a race for clicks, views or shares.
It’s a race for user profile data.
Here’s what this looks like in a few years from now, and why conventional “traffic sources” are becoming irrelevant…
The Money is in the List… Right?
In the near future, this age-old marketing mantra is simultaneously never-truer and also completely irrelevant.
You could say it’s somewhat complicated. Let’s explore why, starting from the beginning.
Just over 10 years ago, when I had to somehow pay rent each month by selling enough of my skateboards online, my options for prospecting new customers were as follows:
* I could try to rank / compete on Google, Yahoo & MSN for popular keywords and wait to be discovered by my next self-qualifying customer (hopefully). Either organically or with PPC.
* I could list products on Ebay to reach consumers already directly in-market, and who might be actively searching for a new board at that time.
* I could buy ads or post specials in niche skateboarding sites/forums.
* I could broadcast something to my email subscribers.
And that’s about it. Of all those traffic strategies, it didn’t take long for me to realize that email marketing was where my “rent salvation” was most likely to be attributed.
Because unlike the other channels (search, marketplace, or media placements) – I wasn’t relying on an average conversion rate from a fixed daily volume of available traffic. Rather, I could “tap” my total accumulated audience on-demand, where my reach was instantaneous and limited only to the size of my list. And to top it off, list traffic converted like no other, and was essentially free… once you built it.
That is why the old mantra rang true for the majority of businesses trying to sell online. Email marketing allowed you to build your own prospect pool, and then push them down the product funnel whenever you wanted.
So, naturally the smart play was to grind out the necessary struggle for traffic in the beginning to build your list as large as possible – after which you only had to occasionally run paid traffic to counteract your attrition (unsubscribes), when necessary. This process took several months, at the least.
Today, email remains a critical CRM channel for most brands. And yes, there’s still plenty of money in the list. But unlike times past… it’s no longer considered the ultimate traffic source.
And that’s not because email was somehow usurped by another traffic channel… but rather because getting traffic is no longer the “hard part”.
Even just a few years ago, driving traffic was all about setting up multiple streams of visibility – where enough individual trickles might just add up to a sizeable flow of activity. (One day).
But today, with the ascendancy of Programmatic & Facebook Ads having aggregated the entire internet & app ecosystem into a centralized inventory pool – literally anyone can walk right up to the internet’s proverbial fire hydrant, clamp on a hose, and drive a massive torrent of traffic – instantly.
And with the right audience data (or data-derived targeting algorithm) to stack on top of that 2+ billion user audience – these are visitors who can be every bit as targeted and “in-market” as the ones you previously had to earn through months of writing content, begging for backlinks, or aggressively competing to acquire (search by search) with Adwords.
In short, list-building is no longer really a traffic source so much as it’s a CRM asset, a targeting tool, and a segment of your overall 1st-party data.
However, the money is still in the “list”, of sorts. But properly utilizing your list for targeted traffic is now a different process.
Because in today’s marketing landscape – where you can snap your fingers and reach 98% of internet users instantly – the challenge is no longer in driving traffic. The challenge is in figuring out how the hell to filter it into profitable audience targets.
Which brings me to the future of traffic…
Show Me Your “One in a Million” Customer…
…And I’ll Find You a Million More. Instantly.
~ Every Self-Respecting Ad Platform in 2 Years
For most people, their marketing data (analytics) is basically just a barometer.
“Oh, looks like traffic is up this week, and sales seem good. Sweet.”
The slightly more serious users view their data as a source of insights about prior events. It’s a way for them to react to things that already happened, and perhaps to inform smarter decisions.
Sophisticated marketers are using their data to create direct ad targeting segments (retargeting) in tandem with their CRM strategy (rules-based marketing automation). Most brands with a half-decent marketing department are utilizing 1st-party audience targeting & CRM, at least to some extent. And basically, this just amplifies the impact of whatever they’re doing to drive first-click traffic (aka: prospecting).
But the really sophisticated marketers are also using their data as a primary traffic source – massively expanding their prospecting reach, and effectively multiplying their TAM (total available market).
Right now today, FB’s Lookalike Audience feature is the best vehicle for doing this. With as little as 100 connected emails / phone numbers, Facebook can build you a targeted audience into the millions. Even in “small” niches. This is truly game-changing.
But it’s not the only game in town… and it sure as hell won’t be the last.
Make no mistake: The traffic source of the future is algorithms.
Algorithms – or if you like, Intelligent Audience Models – that take seed data (yours), and with that “seed”, build massive profile-matched segments across the entire effective internet user base. A user base that’s instantaneously reachable with programmatic advertising across the aggregate publisher networks, including Google & Facebook’s inventory.
And yes, while there’s a few options for audience extension already – these algorithms are in their infancy. They will continually improve, and probably verticalize (become industry-focused). As with Facebook, it’s going to start with the major publishers (who already have access to massive stores of their user’s data), but I believe we’ll see an entire category of Ad Tech emerge in the very near future – where the sole purpose of these platforms will simply be to create various types of targeting logic from your existing data.
Eg. “Go find 1 million prospects that match the profile of my top 100 fastest-converting customers to-date, but who’ve never been exposed to my brand.”
And things will get really exciting when these Audience Models start effectively blending in accurate 3rd-party data.
Eg. “Go find 1 million prospects that match the profile of my top 100 highest-LTV customers to-date, who also earn more than $150K annually, and who actively use their FitBits.”
Yes, the effectiveness thereof depends on the accuracy of this additional targeting data (3rd party) – as well as the participation of the data providers (FitBit, in this hypothetical example) – but as data increasingly becomes more and more of a saleable asset in the programmatic ecosystem, the volume of 3rd-party targeting layers will only increase in abundance & accuracy.
In theory, you can sort of do this already. Most DSP’s integrate 3rd-party data profiles from companies like Exelate, BlueKai, etc. But it’s not really all that effective yet.
This is changing, though – and it won’t be long until the EPC’s from 3rd-party data consistently exceed conventional placements.
Ultimately, what this means is that in the very near future, the vast majority of your traffic generation will be the metaphorical equivalent of Professor X scanning the global populace, and pinpointing all of your most promising prospects (on the planet) in a matter of seconds.
Well… so long as your market exists 😉
So… What Does it All Mean (For You)?
If we’re finally witnessing the fabled era of centralized & unlimited push-button traffic, across every channel & device – traffic that’s far more targeted than ever before…
…is there any opportunity here for “little guys” like us, aside from early adoption?
The opportunity for anyone who’s properly grasping the implications of using instantaneous, algorithmic targeting across the entire connected web (PC’s, mobile, IOT, etc.) that’s seeded by your own customer data (your list)… is truly staggering.
In simple terms, what this means is that your capacity to scale any viable business to critical mass is now no longer limited by time, nor is it limited by access.
Reaching the majority of your entire available market is now something that you could literally accomplish inside of 24 hours. And the scale & precision of which you can do so will only continue increasing, as the various matching algorithms improve, and as the ecosystem of 3rd-party data expands.
As a result, my prediction is that in the early days (the next 2-3 years), early-adopting marketers will find themselves with virtually limitless targeting capabilities. And combined with the sheer scale of inventory – CAC’s that mirror the early days of penny-click Adwords. Good news for the lazy… at least in the short term.
But inevitably, in time, the moat will widen as demand begins to catch up with supply. (Or as the data vendors figure out how to inflate costs relative to targeting – whichever comes first).
So as I see it, here’s the best way to cash in on the emerging bonanza – as well as thriving well into the future as the new world of algorithmic traffic matures…
1) Start Collecting As Much Customer & Conversion Data as Possible.
This goes beyond just dropping cookies (retargeting audiences – which is still definitely something to start doing yesterday, also). You want as many end-points as possible across your entire userbase.
Emails, addresses, phone numbers (especially mobile), Twitter handles, attribution data… everything. You want to do this so that you can start linking your marketing results to real people.
And you want a variety of data-points to increase the likelihood – and accuracy – of audience filters & algorithmic profile-matching.
This is the new “list”.
The more data you have on your customers… the more customers you can find.
(Something, something, metaphor regarding “it takes money to make money”).
2) Start Experimenting With Existing Tech.
Obviously, start using Facebook’s Lookalike capabilities. Do this yesterday. Use different types of accepted data, and see which customer segments drive the best results with FB’s current algo’s.
Then, branch out into the emerging alternatives. Here’s a few to try out:
* AddThis’ Custom Audiences (you’ll need a DSP/DMP to activate)
Just start running traffic, and see what types of data (and parameters) works best for each platform. The sooner you get started, the more of a head-start you’ll have on this as better tech, and better platforms, continue to emerge…
3) Get Positioned to Actually Make the Most of This. ASAP.
In simple terms, here is what you need:
* A scalable business model where you actually collect (and own) your customer data.
Yes, I realize “the Amazon model” is fun and exciting. But you have very limited data assets, and you’re ultimately just a supplier for Amazon. Consider pivoting to an E-comm model under your own brand, domain, traffic channels, etc.
Similarly, affiliate marketing is wonderful for (temporary) cashflow. Well, at least when you manage to hit a home-run. But just like the Amazon model, you’re flying blind in terms of data.
(Not to mention, it’s generally a zero-equity business… but that’s a whole other blog post).
And lastly, ensure that you’re not too niched-up. Your top competitors should have thousands of active customers.
Or alternatively, if you’re an agency or consultant – get clients who match this profile.
* A funnel that turns a profit with paid traffic.
I won’t pull any punches on this. If your business model doesn’t run profitably with paid traffic, then first of all – why are you wasting your time?!?! – and secondly, just figure out how to make it work. Period.
Fix your conversion rate with split-testing and if need be, by hiring copywriting talent. Raise your LTV with upsells, improving retention, testing prices, testing various CRM & marketing automation tactics… etc.
There’s a billion resources and tools out there that will simplify and guide you through this process.
Is it fun? No. It sucks. This is the worst part of the entire process, because you have to essentially be willing to run at a controlled failure-rate, all the while scrambling to solve the “profit puzzle”.
But once you DO turn the corner (and with enough persistence and – chiefly – the willingness to adapt), then your capacity to scale will only be limited by internal bottlenecks.
* A CRM and/or data management platform
Google Analytics won’t cut it, because they won’t let you connect user sessions to PII (identifiable information). Which is essential if you plan on extending your data beyond the walled-garden of Google’s ad eco-system.
You need to be able to create valuable segments from your existing leads & customers – as well as collect & store as many datapoints as possible.
Again, there’s a number of options out there that will fit the bill for this, or at least mostly fit the bill. Examples include Ontraport, InfusionSoft, and even autoresponders like Aweber (to an extent).
The key here is to be able to quickly export customer data by filters like: Biggest spenders, recent converters, customer type… etc.
Bottom line – your list is more than just your list.
In the new age of Hot Models as the traffic source – your “list” is the entire powerplant. And its horsepower is determine not by its size per se, but rather by the amount of end-points you’ve collected, and the creativity of your “seed segments”.
Get good at this now – while this tech is still just an infant – and you’ll be unstoppable by the time it’s fully grown in the next couple years.
I think that’s it for this one.
While I do have somewhat of a dog in this fight (Intelligent Modeling), it’s still too early-stage to introduce at the moment.
So for the meantime, I suppose you’ll just have to settle for my ramblings, sans agenda. (I know, crushing.)
Thoughts on this? Questions?
Are you experimenting with lookalike targeting already?
Are you curious about the airspeed velocity of a laden swallow?
Feel free to post a comment below, and I’ll definitely reply back before 2017.