bayonblog

The online marketing outhouse is full to the shitter’s brim with guerilla online marketers. You know the type, they track their entire operation in a series of confusing spreadsheets, pass keywords straight through to their affiliate networks and generally have no larger idea what they’re doing in architectural or strategic terms, but they make money. These scrappy, rascally nogoodniks are entertaining as Hell and often quite innovative and effective. Why do they make money? Because we’re in online marketing, not online marketing, and these characters are marketers first and foremost.

As they wander our little minefield of a niche, they inevitably bounce off of walls like little market inefficiency-sucking roombas, vacuuming up dollars in one direction until they hit an obstacle, then changing course to repeat the process in another direction. Similar to the wildcatters of yesteryear, they’ll blow out a well and leave it uncapped as long as they can make enough money off the deal to remain profitable - sometimes fucking up things for the rest of us .

Inevitably, though, they begin looking around and asking questions. They want to know how those of us in the professional world know all this crazy stuff about cloaking and keyword-ad-landing page targeting and list management. Basically, they get tired of bouncing off the walls and decide they want to become a slower, but more stable and reliable operation (would you rather be Exxon or my broke nonagenarian neighbor back home with no thumbs?). There are a number of ways to go about this and they’re all worth covering, but guerilla behavioral targeting popped into my brain today because it has been in the news lately and it’s one of my favorite topics. If you need evidence of its ascent popularity among the goatee-stroking weberati, simply google “behavioral targeting” or “YAHOO SmartAds.”

As you’ll see, articles about behavioral targeting are chock full of highbrow terms like “cross-channel” and “inventory databases” that try to astound and amaze the reader. Online marketers, generally speaking, are horrible about doing this. We’re marketers and we don’t stop being marketers when we’re not selling products directly, so what do we do when we want to tell people about a new concept? We market it instead of informing about it. Honest to God, if tomorrow every man could suck his own dick, online marketers would be the first people to post on the internet, in glowing terms, about how amazing their autofellatio experience was while imploring you to buy their eBook on how to suck your own dick. As stupid as that sounds, two days later they’d be caught trying to sell eBooks to one another about how to make money on the internet selling eBooks about sucking one’s own dick (if you don’t believe the amount of internal stupidity in the OM community concerning eBooks, visit The Warrior Forums). If you can imagine, it actually would go downhill from there.

Back on topic, in order to make the concept of behavioral targeting accessible and applicable to the guerilla operator looking for a little order in his or her universe, let’s strip out the glittery hoohaa and mojo in order to emerge with a few kernels of shit-stained, corn-like wisdom.

First, what is behavioral targeting? Well it sounds fancy. Is this like marketing spitoons to rednecks? Given a demonstrated preference it could be. Essentially, behavioral targeting is deciding what your user might probably want to buy based on some demonstrated action and then trying to sell it to them. In the real world, where we don’t need fancy new terms to describe ancient concepts, we see behavioral targeting every day. If you go to the same bar and order Singha, a Thai lager, for a couple of days, the bartender is going to notice this and, on the third day, he or she will probably offer you a Singha or suggest another kind of lager. Sim sim sala bim, you have just been behaviorally targeted by your local bartender! Fucking magic! YAHOO: Catching up with bartenders everywhere since 2007. Let’s analyze the factors at work here though, just to make sure they’re obvious:

1. You demonstrated a preference for a certain brand of beer: Singha.
2. You demonstrated a preference for a certain type of beer: lager.
3. The bartender was making a mental record of these preferences.
4. After two visits, a given level of probability was reached in the bartender’s mind concerning your beer preference.
5. On the subsequent visit, the bartender used his or her knowledge about your demonstrated preference for Singha and your probable preference for lager and made a suggestion.
6. YAHOO sends in teams to model this complex sociological interaction and capture lightning in a bottle - Shazam!

This kind of dynamic targeting has been going on online forever, at least since outfits like firefly started working on it back in the heady dotconomy days. The most famous online example, of course, is Amazon’s “we thought you might like” content that consists of books or other products suggested based on your browsing habits, but they didn’t invent the concept. What makes it interesting to apply in online marketing is the distributed nature of this field. We operate through many different promotional channels (PPC, email, blogs, SEO, etc) using many different revenue generating content sources (CPC, affiliate offers, etc); However, unlike Amazon, we often control only some or very little of these channels, making the collection of information more difficult.

Because of these factors, in the guerilla online marketing world it’s quite common that you don’t have more than one shot to hook a user, let alone guess at his or her preference. Let’s suppose that you’re advertising weight loss affiliate offers through a PPC campaign that dumps users on a landing page that you control. A user clicks on your PPC weight loss ad and lands on your landing page that has several different offers, one diet program, one fat loss cream and one abs of titanium machine - all of which are form submits. The user enters information into your hosted diet program form and presses submit - hallelujhah! You then record this info into your ESP (your email service provider, like getresponse or aweber) and pass the user on to the affiliate network who then passes the user on to the diet program offer. Unfortunately, unlike Amazon, the probability of that user visiting your page again naturally or even clicking on another of your PPC weight loss ads at a later date is probably very small, so you have to gather what info you can when you can - and what info was there to gather?

1. The user demonstrated a preference for a certain market vertical: weight loss.
2. The user demonstrated a preference for a certain type of product in that vertical: diet programs.
3. You collected the user’s information.

So far so good, but if the flow of information stops there then you are officially worse at marketing than your local bartender. Dumbass! Obviously we’re assuming here that you, the guerilla online marketer, don’t have the capability to create a database to collect this information automatically and the programming skills to create systems that act on this information dynamically, so what can you do to behaviorally target this user? Well, keeping it simple, there are a couple of things.

You can start at the ESP level. Using the current example (where you’re hosting offers and passing the data) and pretending that you’re using aweber, you could add an autoresponder labeled “weight loss” and define custom fields to hold the demographic information you’ve captured:

1. Create a custom field in your new weight loss autoresponder called: subvertical - (diet, exercise, etc).
2. Create a custom field in your new weight loss autoresponder called: keyword - (’how to lose weight’, ‘herbalife’, etc.)
3. Create a custom field in your new weight loss autoresponder called: referrer - (http://www.google.com/… = ‘Google’, etc).
4. Create a custom field in your new weight loss autoresponder called: source - (the proper name of your site - ‘Super Diet Site’, etc).

Now, change your normal hosted form to point to your aweber autoresponder form. Insert all of these variables automatically using either hardcoded variables <input type=”hidden” name=”subvertical” value=”diet” /> or dynamic variables that you collected from the incoming PPC link <input type=”hidden” name=”keyword” value=”<?php echo $keyword; ?>” />. Also, be sure that you have the meta_redirect_onlist hidden form field set, in addition to your standard thank you page, to keep your users from seeing an error message if their address is already on the list. The final step in ensuring the user flow is to verify that your thank you pages point to either a cloaked URL on your site (that forwards on) or to your appointed affiliate offers. If you’re able to pass information to these pages to pre-populate forms on the affiliate end you can still do this by using forwarded variables (aweber’s terminology) in the appropriate part of the querystring so that they get transmitted back to your cloaking page or directly to the affiliate network (this might look like &em={!email}).

Once you’ve done this, go to your affiliate networks and pick out one, or a few weight loss related offers and build an email around the offer or offers. This can be a welcome message or a pure sales message or whatever you feel is most appropriate to your campaign (this depends heavily on whether you’re running a branded site or not). This email will be launched automatically when someone joins your list. Don’t forget in creating this email that you have available to you a number of variables now, including the keyword that the user searched on {!custom keyword}, the network they searched for this word on {! custom referrer} and more.

Having created your initial follow-up message, it’s time to schedule some follow up messages at regular intervals - once a week, once a month, whatever you like. It’s also time to segment your list further using the subvertical custom variable and the ‘view’ functionality of aweber. You’ve stored ‘diet’ in the custom subvertical field for every person that filled out the diet program form and ‘exercise’ in the custom subvertical field for every person that filled out the abs of titanium form. In the Leads section of aweber, search in this autoresponder list for leads where the subvertical is ‘diet’. When the list is returned, save it as a view. Now, go back to the messages section and build an email marketing message specifically about diet programs. Do the same for all your other subverticals and, when you’re ready, send them to your newly saved views that are made up entirely of leads that have expressed an interest in the subvertical that the email message is built for.

Assuming you’ve done all this, what have you now accomplished? You’ve got a cross-channel, behavioral targeting system that sees an incoming PPC click, sends information about the PPC campaign to your ESP, passes the user on to an affiliate offer and then automatically launches an email campaign for each new registration based on the market vertical (weight loss) of the product that they signed up under. This system will email them at scheduled intervals with your pre-selected offers and automatically create revenue with no further intervention from you. Furthermore, you have the same users targeted at the subvertical level (diet, exercise, gels & creams) by view and at the click of a button you can send them messages specifically targeted at that subvertical.

Obviously I skipped over a lot of the mundane details here under the assumption that anyone reading this already understands their ESP program, their PPC program and how to build their own landing pages. Sure, this system wouldn’t pass muster with the standards and practices crowd and it’s nothing that one would use at the enterprise level because there would larger systems available to gather more specific data, access internally controlled channels and carry out more specific actions, but if you’re a guerilla marketer - usually a one man shop - you simply don’t have these capabilities at your disposal. Assuming you’re one of the guerilla gorillas and you’re using a cobbled together system made up of third party tools, this is a quick and easy way to get your systems working for you to automatically generate revenue based off of user behaviors.

This entry was posted on Sunday, July 22nd, 2007 at 11:48 am and is filed under Marketing Tech Concepts. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.