Closing at Volume – Using Not Abusing Purchase Based Audiences

In this third article on audience targeting, we’ll take a look at purchase based audiences, glorified, high efficiency, sometimes misunderstood, and often misused.  Purchase intent, also known as in-market audiences are the most granular tool marketers employ, for the sole purpose of driving immediate sales and revenue.

What is Purchase Intent?

Quite simply, purchase intent is the likelihood of a customer to buy a product in the near future.  It’s a no brainer to grasp the concept, but another thing entirely to reliably identify for any given consumer.  To bring the concept back to the prior two articles in this series, demographics are a broad, awareness based strategy, what marketers might refer to as a top of funnel strategy; psychographics on the other hand are a mid-funnel strategy where marketers drive consumers from awareness to thinking favorably of the brand and even having a preference for it.  And at the end of all that effort then is the bottom of the funnel, when consumers are finally making that all important decision on what to actually buy, hopefully with the marketers’ brand in consideration.

Targeting Media by Purchase Intent

When it comes to purchase based audiences, the data usually depends on the length of the purchase cycle.  For products that are low priced and consumed daily – think consumer packaged goods like food, cleaning supplies, beauty products, and the like, the upstream data source is usually some kind of historical database, meaning the audience is really people who purchased in the past.  For higher priced or less frequent products like auto, travel, insurance, or big ticket electronics, more predictive data is typically employed.  Consumers haven’t purchased yet, but they exhibit the signs of nearing a purchase in the near future.  Each technique, historic or predictive, naturally relies on different upstream data sources, and because of that, the data brokers for purchase based audiences tend to specialize in one type or the other.

For historical data, the data brokers and suppliers are typically based in the offline world, and source their data from offline sources.  Long before digital marketing even existed, enterprise data companies like Acxiom and Epsilon offered highly granular and specific data segments that allowed direct mail marketers target specific messages and special offers to highly qualified customers.  Those companies worked with major corporations like grocery stores, retail chains, as well as large public data sources to aggregate and sell lists of customers with the right behaviors, and today they’ve simply extended that model to the web, and bring their offline data online with match providers.

Since the upstream data suppliers for historically based purchase audiences are transaction level databases, and are diligently scrubbed and matched back to specific households by established data companies that operate across multiple media, these audiences can often be cut and customized to an astonishing level.  Looking for consumers that favor one brand over another – say to target your competition’s customers?  No problem.  Or what about getting audiences that are also in the top quartile of consumers for a given product, or even those that shop specifically at one store or another?  All is possible through the magic of enterprise data management systems, the issue is more around scale and cost effectiveness than what’s possible.

Predictive models on the other hand, focus more on observable actions, like seeing a consumer view lots of pages on the same type of product, say a plasma screen television, or maybe requesting a quote for a particular type of car from a local dealer.  Some purchase based audiences might even get defined with highly correlated behaviors, like if a consumer buys a plane ticket from New York to San Francisco, hotels and car rental companies in San Francisco might very well consider that person in-market for their products.  The trick with predictive audiences however is to understand the recency of the data, that is, how long ago a consumer could have demonstrated that behavior before they expire out of the audience, the reason being that after a period of time, the consumer will have made their decision, and purchased something.  There’s no sense in targeting hotel ads to a person two weeks after their trip, so it’s important to know how predictive audiences are defined.

Everything in Moderation…

The critical thing to understand is that purchase intent strategies are largely the culmination of all that other work, just one part of a whole, and not much of a solution on their own.  Consumers don’t just get to the bottom of the funnel on their own, they’re influenced along the way, in each part of the purchase funnel with different messages.  While purchase based audiences no doubt valuable, marketers are often at risk of being seduced by the easy, direct response style optimization they can do at the bottom of the funnel, and can start to ignore the effort they put into getting those opportunities setup in the first place.  The classic example is the fallacy of last-touch attribution models, which tend to focus spending, wrongly so, on the bottom of the funnel, rather than trying to optimize the balance of resources against each technique.

Purchase based marketing and optimization might drive the most efficiency, but a comprehensive strategy that works all levels of the purchase funnel will maximize the volume of sales, not just conversion efficiency.

Demographics Explain Who, but Psychographics Explain Why

In the first piece in this series, we looked at demographic information, and how marketers use information like age, gender, and income to tailor their advertising strategies for greater effect.  But even the most sophisticated demographic target is still broad, so how do marketers take the next step and start to zero in on the right customers to drive sales?  The answer is usually found in the complex world of psychographics, where advertising and psychology merge to produce the granular, qualitative information to understand what really makes a customer buy.

What are Psychographics?

At their core, demographics describe who people are, and psychographics describe what they care about.  From a marketing perspective however, psychographics are usually used to explain why people buy a product, and the attitudes, opinions, and personality traits that drive them toward a product.  Psychographics are inherently more abstract than demographics because they are multi-dimensional by nature, and cover subtle elements of what makes a person tick.

Today, thanks to an explosion in communications technology, consumers are more fragmented than ever before, so understanding and marketing based on psychographic qualities couldn’t be more important.  From a consumer point of view, the internet has made it easier than ever to discover subcultures, connect with distinctive communities that share the same unique interests, and curate their world.  You can see marketers respond to this reality by adopting wider identities, and narrower marketing strategies.

As an example, consider athletic shoes – from a demographic point of view, a broad range of people are in the market for the product, from young to old, both men and women, for all kinds of reasons.  From a psychographic perspective though, some customers might care most about performance, while others concern themselves with the fashion appeal of the product, while still others just want a particular brand as a status symbol.  For a shoe company to maximize sales, it needs to understand these trends to design the right products, and talk about them the right way.

You need only look at major shoe retailers to see the model in action – endorsements from celebrity athletes for the performance crowd, options to customize the colors and materials for the fashion crowd, and rare, limited editions for the status conscious crowd.  From an advertising perspective, the brand talks to one crowd very different than another, so the ads on ESPN are very different than on SneakerNews.com.  There’s no doubt that shoe companies fully understand the demographic qualities of each customer base, but in many ways, the psychographic elements are what really drive the products, positioning, and sales.

Targeting Media by Psychographics

Unfortunately for advertisers, psychographics aren’t nearly as easy to deploy in targeting strategies as demographics.  For one, psychographics aren’t easy to standardize on a taxonomy – brands won’t necessarily agree what qualifies someone as a ‘sneakerhead’ the way they will ‘Age 18 – 24’.  It’s also far more difficult to collect the information from people in the first place – while demographic elements are widely available from a variety of public and private sources where someone might fill out a form or register, but the same can’t be said for psychographic data.  In most cases, there isn’t a pure data-driven signal, but rather a combination of data and context that works as the best proxy.

Going back to our example of sneakerheads, aside from a passion for collecting rare sneakers, marketers might learn that these customers also have a strong affinity for hip hop music, are more likely to watch professional basketball more than any other sport, and buy a new pair of shoes once a month.  Having gathered this information, marketers can now combine various behavioral elements in context to reach their psychographic target.  Instead of just advertising on whatever sites attract their demographic, or focus exclusively on sneakerhead content, the marketers might start to look at sites that focus on hip hop music, or professional basketball content sites.  They might target people who they’ve tracked on both basketball and hip hop music sites with ads, perhaps email every user who last purchased three weeks ago with a special offer to capture their next purchase, or start to integrate those cultural topics and elements into their social media strategies.

Over time, marketers can start to build a truly sophisticated profile and targeting strategy as they get more and more efficient at driving sales that might understand what specific basketball teams, what specific hip hop artists draw the most interest from the customer base, and even start to segment the different trends between heavy buyers versus light buyers, and what makes someone move from the latter group to the former.  While far more complex in nature to first understand and then execute in the marketplace, adding psychographic elements to a targeting strategy offers a greater payoff when deployed effectively.

Everyone’s day has a rhythm. Surely yours does too. Maybe it’s come in early, answer some emails before attending a couple of meetings.  You squeeze in some heads down work before heading to lunch.  After lunch it’s catch up on the morning’s emails and so on…

What would happen if you answered emails instead of mid morning meetings? Would response times to potential clients be faster? Would faster response times mean quicker decisions? To find out; give it a try.

Spend a week approaching your day with the opposite schedule as normal.  You’ll start to notice differences.  The faster you respond to emails, they faster those people will respond to you.  Return voicemails as soon as you see there is a message, rather than the end of the day.

Take the time to give this strategy a try and I can promise it will lead to quicker decisions and better relationships.  Don’t believe me? Apply some analytics to the experiment.  Plugins like Xobni will chronicle your emails to tell you who you email most, who you respond to fastest and who responds fastest to you down to the second.  Compare the before and after of your experimental weeks to see what kind of difference was made.

2)   Never drop the ball, really, never

Every now and then we all get an email follow up 3 months late saying “this one must have fallen through the cracks…” Maybe you’ve even sent a few. Perhaps you came across one of these conversations and realized you had dropped the ball. In taking the more passive approach, the conversation was ignored to save face.  Every conversation that’s left behind is a lead left behind.

Try to avoid leaving leads behind by responding to every email.  Every one: Calendar reminders (Yes, I’ll be there. Who else is going?) , spam mail (Sounds like a really interesting product! I’m not in the market right now, but what other products does your company sell?), the messages that explicitly end with “do not reply to this message” and most importantly conversations with other people.  Try it. It’s a great exercise in creating conversation.  More importantly, you can look through your inbox and see that you sent the last message in every conversation.  Presumably, this means the burden of response falls on the other participants in the conversation, ensuring you haven’t dropped the ball.

Just responding to every email sounds easier than it actually is. Responding to every email doesn’t mean babbling nonsense that adds no value to your conversations. People will stop talking with you if you do.  Extending conversations beyond the initial question is an art and a good way to practice this art is by using the “yes, and” tactic from improv comedy.

Using the “yes and” strategy will challenge you to add value to every email you send.  There are thousands of ways to do this from asking for an opinion, linking your response to a blog post that takes a position on the conversation or including links to examples that expand on the topic.

3)   Don’t email

Strategy 1 was to change your schedule so that you could respond more quickly when people are email you. Even with that change, there will be some people who can’t be reached through email.

Personally, it drives me crazy when you’ve emailed someone who you find Tweeting, Facebooking or worse yet, playing “hippopotamus” on Words with Friends before responding to your email.  Is Words with Friends that much more important than my email?!?

Perhaps. But, perhaps not. I have some contacts who’s email inboxes are such as mess that I’m lucky to get a response once a month regardless of my use of CAPSLOCK in the subject line.  But they’ll respond to a Tweet in seconds.

Spend a week finding different ways to follow up with the people who contact you through email.  This will surely drive some of your contacts crazy. When it does, stop and communicate with them they way that doesn’t drive them crazy. Otherwise, find a different way to respond to each email you get.

Someone sends a 200+ word email? Pick up the phone and call them back to talk through the issue.  It’s probably more quickly resolved on the phone anyway.

Someone emails asking for examples of sites in your network? Collect the examples and bundle them with a bit.ly and send them in a Tweet. You may find that much like my contact, Tweets rise to the top of other’s work flow faster than email.

This approach is beneficial for two reasons:

First, you’ll find that you can get responses more quickly from certain contacts by using different social channels. This can be beneficial now, as well as in future conversations that may be more time sensitive.

Second, you’re forced to think more creatively.  Email is a fairly universal means of communication in today’s business world, but is it really the best way to communicate your current message? Rather than sending an email with 4 attachments could the information be screenshot and uploaded to Pinterest? When you take the time to find ways to communicate beyond email, your customers will notice.

Options are sometimes good for consumers.  Increased competition lowers pricing and encourages innovation through market forces.  A number of options can sometimes be overwhelming though.  Suppose a business decides to do search engine marketing instead of display, social or pre roll advertising.  Now that business is faced with having to choose which of the four above search engines will produce the best return for them, and we didn’t even bring up AOL search.

As an ad network, the online advertising options likely overwhelm many of your prospects, businesses and agencies alike.  So how do you, as a network, demonstrate value over pay-per-click platforms and explain the difference between your network and competitors?

The answer is two fold.

First, make the decision making process easier for potential customers.  Any smart potential customer is going to talk with you and your competitors, so why not facilitate the comparison? People love visuals (hooray infographics?) so help them align their decision making visually.  Find The Best aligns things nicely between ad networks.

Make comparing yourself and other networks easy for buyers. Not only will you make potential customers lives easier, you have the chance to define the decision making process and highlight your strengths. It’s always easier to write on a clean slate than it is to erase.

The second strategy works well in some instances and poorly in others. It does require you to know a bit about ad agencies though.  Agencies generally get paid in a couple of ways when it comes to media. The first, development of creative is where agencies make the majority of their revenue.  The second way, is media buying fees which are consistently being pushed down to the point of being marginalized.

For that reason, creative is the driving force behind the agency’s (your customers) success. The agency has to demonstrate value to its customers by developing creative and then buying and displaying that creative, while really only being paid for one of these two activities.  The one activity they are being paid for is creative. With that said, how can you make your customer, the agency, look like a rock star?

The answer may very well lie in more creative. Does your client have a target audience? Could they reach it more effectively if they segmented that audience? Because segmented audiences need segmented creative. Can you help them segment?

If so, everyone wins. Case studies are a fantastic way to do this.  While it would be nice to think that all decisions are made based on qualitative data, the reality is that story telling is often more powerful than numbers alone. So don’t use a terribly undefined set of statistics that really provides no value like the information below on why Mom’s like smartphones from Blogher.

But, actually segment the audience with data , case studies and anecdotes to reinforce your points and the benefits of segmenting on your platform. Again, to emphasize, this doesn’t include data from ComScore or Alexa. While that data is useful, it’s available to everyone.

To effectively build and demonstrate the ability to target an audience and convince an agency to convince it’s client, segmented audiences and creative is a worthwhile investment (cause that’s really how it works), will be hard.  Maybe even some concessions on the part of your network will be required at first.  But it will be worthwhile in the end as your network ends up with a case study, the agency gets more creative and the client gets a better preforming campaign.

Very few other networks are doing this for clients. In fact, I had a hard time finding good case studies to use as examples for this post. Doing the things that no one else will do is exactly what will win you customers. This starts with case studies.  Who is your best current client? Would they try multiple creative and experiment to target a segmented audience and allow you to use that data? You may have to comp the cost of the impressions, but that case study will go to every potential client for the next year.  Is that a worthwhile investment in a differentiator for your business?

One amazing advantage the smart phone has that the PC doesn’t have, is an integrated feature built right into every operating system – through it you can have access to a real-time, digitally streaming, two-way communication devise, delivering instant engagement between an advertiser and target customer.  No it’s not some fancy new mobile-cookie-retargeting-reverse-geo-fencing rich media chat-based ad unit, but a good old fashion phone call!  Advertisers, especially small businesses, love it when their phone rings, and on the other end is a customer saying “Hi, I’m interested in this product you have to offer”.  With a conversation like that it’s no wonder a phone call still converts higher than any search campaign and it’s a technology every business already supports today.

So if you think about it from that perspective phone calls should be counted as a form of lead and optimized to – just like any other online engagement or CPA event.  (BTW – if you think you’re in the clear from CPA by charging CPM or CPC – think again!  All smart money advertisers are constantly backing out every dime spend on advertising to an ROI – in most cases its CPA or CPL – Cost Per Lead)  In the early days of display, we almost felt like showing a phone number either in the ad or on the website was cheating or a cop out.  We’re here to do “online” advertising after all so if we can’t market our advertisers products effectively enough where the consumer feels they have to pick up the “offline” phone to call the business directly for assistance, we’ve failed in some way.  Plus, even if you wanted to, how would you know how many phone calls your ads produced?  If an advertiser said, “I want my CPA to be when my phone rings” how would you know which ones to take credit for?  Here’s how:

Tracking phone calls is actually remarkably easy.  I know some of you ad ops folks out there probably figured, “OK, I’ll setup some click event on the display ad to ping a tracking pixel every time someone taps the “call now” button”.  OK, that’s great but what if the user calls the phone number without clicking on your ad?  How do you know how long the call lasted for?  Did it connect?  Did the business pick up or did it go to voicemail?  Was it the new customer that the advertiser was hoping would call them or an existing customer?

The solution is actually built right into the phone communications industry.  You can setup what’s called a “Call Tracking Number” or CTN for short. The CTN is nothing more than a new phone number (ideally as close as possible to the advertisers original phone number) and anyone that calls the phone number is just forwarded to the advertisers original phone number.  Think of it like a click redirect URL.  The advertiser has one phone line, but multiple phone numbers floating out there where at least one of them (your CTN) is sharing a rich set of usage results and performance metrics back to you.  Just like a phone bill, you can track everything from caller ID, call duration, call status (did it answer, go to a busy signal, or ring forever?), call type (mobile/cell, VOIP or landline).  With the advertiser’s permission, you can also setup call recording so you or the advertiser can go back and asses call/lead quality.  Did this call convert into a sale?

There’s also plenty of drawbacks to this approach you should be aware of as well.  We all get random “wrong number” calls.  Ever get a new phone number and for a couple months you keep getting calls from people asking for “Bryan”?  (“What you mean this isn’t Bryan’s phone number any more? Do you know Bryan’s new phone number?” No I don’t!  Otherwise I’d call Brian and give him his old phone number back! Stop calling me!).  Adding CTN’s increases the risk of this happening to your advertiser, so it’s a delicate balance.  There’s also a finite number of phone numbers out there (unlike click redirects) so depending on how picky the advertiser is about there phone number, it might be a challenge to find one they are willing to use.

There’s some more to this that I’ll try to expand on in future posts, but at the end of the campaign, imagine sending a report back to your advertiser saying – “As a result of our ad campaign, here’s how many people were interested enough in your product that they called your business and talked to your sales rep(s) for this average length of time”.

Just think about how compelling that is as a value statement.

Traditionally, advertising has been all about content adjacency – marketers tended to place their ads with publishers who had high-quality, professional content that attracted a certain type of consumer likely to buy the marketer’s product.  Cookware products advertised in food magazines, sneaker companies advertised during sports programs, banks put ads in the business and finance section of local newspapers. In short, publishers attracted and aggregated large audiences, and marketers bought media with those publishers, who acted as proxies to reach a desired audience.  Increasingly however, marketers are going around contextual signals and using data to reach the exact type of consumer they want, regardless of what kind of content they’re on.

Whether this is a good or bad thing remains to be seen – for the sake of this series, we’re just focused on explaining the how & why of what’s happening, and there’s no denying the shift in buying tactics – data is here to stay.  As an initial primer, let’s start with consumer attributes 101: demographics.

What are Demographics?

Demographics are the bread and butter of audience based targeting, not to mention old friends to marketers, who have been thinking and researching consumer demographics for decades now.  Demographics are statistics, inarguable facts that describe a person, and this is important because you’ll find that many behavioral elements used in targeting like psychographics, lifestyle, or purchase intent characteristics aren’t necessarily known, they’re inferred, or guessed at with a certain degree of confidence, but we’ll get into that in the next part of this series.

Demographics cover black and white facts – your age, gender, ethnicity, income, if you’re married, if you have children, if you own a pet, if you own a house, things like that. Demographic attributes can change over time, but they are mutually exclusive qualities – you can’t be 25 years old and 35 years old at the same time.  When marketers can understand these facts across their entire consumer base, they can learn how to drive efficiencies with their messages. For example, cereal companies understand that while pretty much everyone buys cereal, the people that consume the most product and buy the most often are married women who have at least two children between the ages of five and sixteen.  Similarly, a sports car maker might know that single men age 25 – 35 are the most interested and receptive to their ads, but that most of the people who actually buy their product are married men age 45 – 55, who also make more than $250,000 a year.

With the facts in hand about their consumer base, marketers can better plan their ad campaigns, so they reach the most important consumers.  In traditional channels, this often means relying on 3^rd party research which measures the demographic profiles of who watches a particular crime drama or sitcom, but online, marketers can not only buy media based on research, they can pick and choose the exact demographic qualities for all their media, effectively reaching just the consumers they want, regardless of where they are.  For example, audience buying online allows advertisers like the cereal marker to reach just the women with children, regardless of where they are – news sites, finance sites, entertainment sites, anywhere.

Targeting Media by Demographics

So how do advertisers zero in on just the people with a certain demographic profile, and why can’t they do the same thing offline?  There are a few ways to get there, but most of the time, marketers can leverage data that users voluntarily provide, either directly or indirectly, to publishers that offer useful tools and services to consumers.  The publishers then place a cookie on those users’ machines and tie the unique IDs of those cookies to a database where they store the demographic information about that user, separate from the user’s name, or other personally identifiable information.

Unlike other forms of media, the cookie allows the publisher to identify the same user each time they visit the publisher’s website, which allows the publisher to associate information they collected on that user in the past to their current session, and make a decision, like what kind of ad to serve to that user when they navigate through a given website.  For instance, large portals that offer email services, usually require data such as gender and age, and might ask for many more characteristics, when users opt in to use the service.  After the publisher accrues this data, they can aggregate all the cookies which they know are men, for example, and target media to those cookies, no matter where they are in their network.  Of course, not very many publishers have the luxury of these proprietary data assets at any kind of scale, or if they do, aren’t in the advertising business, so they, like many other marketers, rely on data aggregators and data brokers to provide the information.

Data aggregators and data brokers are technology companies that collect and aggregate anonymous user information like demographic qualities from many publishers simultaneous that either don’t have the desire or the resources to monetize that data themselves, and then sell it to marketers and other publishers that can use it.  These publishers are usually service based sites instead of content based sites – online stores where a user has to register to buy something, dating sites where a user has to register to create a profile, or perhaps social media sites where users have to register to use the service.  There are millions of potential collection points, many of which too small to build a dedicated business on demographic data on their own, but which are happy to collect an easy payment from the data broker for providing anonymized access to their user data.  Data companies gain access to this data by placing a cookie on the user when they are on these sites through a small piece of code that forces the user’s browser to call the data company as they navigate on the publisher’s site.  With that cookie in place, the data company can deploy that data to customers by having a similar script on their customer’s websites, which provides the demographic data to ad serving technologies.

It may sound complex, but it’s actually quite simple, and there are very large companies that provide demographic data as a valuable service to the online marketing industry, which allow marketers and publishers alike to better understand users, tailor relevant messages, and drive better and more efficient results through the ecosystem.

Don’t go to these shows without a purpose

I got a lot more out of the San Francisco show than I did from New York because of the time spent ahead of the event lining up meetings.  My purpose in going to these shows is to schedule meetings with people interested in our product so that they have the ability to put a face with a name.  These shows also act as a great networking event that allows you to get in front of people you would not normally be in front of on a daily basis.  Taking advantage of these opportunities is extremely valuable not only for me, but also for the company I work for.

I learned from the New York show that it is probably not worth going if you do not line up at least 5 or 6 productive meetings that all build toward your purpose.  For the San Francisco show, I decided only two weeks prior to the event that I would be attending because I had six great meetings lined up with a couple more tentative ones.  This had me completely booked the Tuesday of the show, but also allowed me to be as efficient as I needed to be.  The San Francisco show also left me with some great connections that may not necessarily be considered sales opportunities, but connections that will be helpful in the future.

Take advantage of the exhibit hall, but don’t rely on it

There are some great companies that have booths on the floor and it is always good to see what is out there, but make sure you don’t make that your main reason for going to the show because you will leave thoroughly disappointed.  You may find some additional companies that you can work with or partner with, but keep in mind, they have a purpose too.  They purchased a booth to increase their business and awareness so you are most likely speaking with a sales or business development rep from that company.  I am not saying this is a bad thing, however, know that going into the conversation.  They will want to know what is in it for them and how it can work toward the purpose of them being there.  Therefore, if you are trying to work with them and put a bug in their ear about your services, you may have to network with them first and get introductions to other people in their company later.  This type of situation can easily come to fruition, but if this is the only reason you are going, you may be let down.

My favorite part of the exhibit hall is the start up section.  There are actually some established companies sprinkled in there along with the start ups so spend some time in that section of the exhibit hall and see what else there is.  Some of the best conversations I had at both shows were at these booths.  They may look small and not worth going over to, however, those are the ones that you will find that may be best suited to partner with you or work with you.  Of course, I may be a little biased coming from a startup myself, but I think seeing the innovation in the exhibit hall far outweighs the established companies because more often then not this is where you’ll spot the cutting edge trends in the industry (not to mention all the energy from the startups).

Find parties that fit your purpose to attend in advance

Something that I did not plan to do at both shows were attend the networking events and parties after the conference hours.  I was taken to a smaller one and realized that there are even bigger networking opportunities at these events than I had originally thought.  I made some great connections in the industry and outside the industry that I was not expecting.  There were a couple of parties at the San Francisco show that would have fit my purpose perfectly, but I got the invite too late and was not able to attend. The next show I go to I will be planning on attending some after party events that allow for even more networking.

Bring sneakers with you

One of the most foolish things I did was not bring sneakers to San Francisco.  I figured I was going to be at the show most of the time and not do much else.  I could not have been more wrong.  Many of the meetings you schedule will be at the event but it’s also common to visit the offices of some of the local companies. My feet were not happy with my lack of footware planning.

Overall, these conferences can be a great experience and help you with what you are trying to accomplish with your job or company.  The only word of advice I will give is to make sure you have a plan and you are making those plans to accomplish a goal.  If you are just going to see a new city, then you may as well just plan a family vacation and get left behind by the people that made the most of the three days at ad:tech.

I would love to hear about some other great or not so great experiences at ad:tech, so add your thoughts to the comments.  I also plan on attending ad:tech in New York this fall, so if you are looking to schedule a meaningful meeting or want to network, let me know and we’ll get something scheduled.

Photo courtesy of affiliatesummit: http://www.flickr.com/photos/affsum/5146419905/

Writing & Publishing Your Digital Ad Spec

In part one, we talked about what an ad spec is, why it’s necessary for digital publishers both large and small, and how to get started writing one by involving key internal stakeholders.  In this final part, we’ll get to the specifics, and how to take an ad spec from concept to specifics.

Getting In the Weeds

Once you cover off on the basics, it’s time for Ops to really dig in, and start testing various configurations to ensure the ad spec works.  That means understanding how to protect against large discrepancies that can seriously effect customer invoices, discovering the outlying factors that can break ads on the site depending on specific site section code, and understanding how to keep the company’s advertising in compliance with privacy policies, and industry best practices.

For example, Ad Ops should understand cache-busting requirements for 3^rd party technologies, and the intricacies of click tracking to minimize the chance for large discrepancies, and so team members know how to QA, and if necessary, correct problematic ads.  For rich media technologies to function, publishers often have to include so-called bridge or gateway files in a local server directory which allows expanding ads to bust, or expand out from iframes and over site content, though these techniques may not always work for all technologies for asynchronous ad calls, or ad slots called with JavaScript instead of through an iframe.  Ops has to understand where the site code can potentially interfere with successful ad executions and consider those factors as they define the ad spec.

Finally, Ops should keep up to date with industry trends best practices as defined by trade groups like the IAB, DMA, DAA, and NAI to keep the ad spec current.  In years past, the IAB started certifying 3^rd party tracking mechanisms to comply with a so-called bots & spiders list, which impacts publishers and means that non-human traffic isn’t counted in ad server reporting.  More recently, trade groups have defined self-regulatory policies for transparency around user tracking, and acceptable 3^rd party opt-out mechanisms.  Being aware of these kinds of services and developments may or may not impact the ad spec directly, but can often inform the necessary diligence the Ad Ops group requires to approve new partners, audit existing vendors, and maintain a professional reputation in the marketplace.

Make Your Ad Spec Darwin-Friendly

Like most everything in digital advertising, change is inevitable, and the ad spec is no exception.  The best Ad Ops teams think of their ad spec as a living document, and one that must regularly evolve.  A few years ago, many publishers didn’t have a mobile business, and certainly didn’t have a tablet business; today however, you’ll find ad specs with detailed sections for each.  Similarly, outdated information can be retired off the spec – it may have been all the rage in 1998, but who really sells the 468×60 ad format today?  Publishers have to update their ad spec to conform with updates to Flash, contemplate new technologies like ad verification services, and regularly update to keep their ad specs current in the marketplace.

This is a good thing, because it pushes Ad Ops to plan for the future, and stay solution oriented.  When faced with a new technology, ad format, or vendor functionality, Ops should ask how to add it to the spec rather than if it should be added.  Plan a semi-annual review at minimum of your ad spec, and even better, quarterly review with internal stakeholders.  In many cases your Ops team might find that they need to involve other teams to update the ad spec.  For example, to enable mobile rich media units, Ops may need to work with IT to update the application code with a vendor’s SDK, and partner with IT to test and ensure that the technology functions correctly before rolling out the update.

Outside of pure tech, sales may have feedback that customers are increasingly asking for new creative sizes, get a certain type of reporting breakout, or target based on a dynamic target, like weather patterns.  Ops will have to figure out how to execute against these requests and how the ad spec needs to change based on the particulars of the product dependencies.  Things are constantly in flux, so regular reviews of the ad spec can help keep Ops on the overall company’s roadmap, both on the technology side and product side.

Ad Specs in the Wild

As much as there is to say about having a well thought out strategy for your ad spec, it’s also helpful to look at some examples.  Thankfully there are a number of companies that publicly post their digital ad specs online, which incidentally, is highly recommended to simplify the communication between internal and external teams.

To list a few organizations in particular, The Washington Post, Yahoo!, AOL, and CBS Interactive all have comprehensive, cross-platform ad specs that are best-in-class examples.  In addition to those companies, the IAB defines industry-wide creative guidelines that can help inform the necessary decisions to make in writing your ad spec.

If you’re a large organization, a well thought out ad spec keeps your business scalable, enabling even the most junior team members to respond to detailed technical questions and requests from the sales organization, which means decision making stays fast and distributed, even at high volumes.  For small organizations, thinking through ad spec requirements is practically a right of passage in becoming a serious digital publisher, a landmark task that forces your company to move from ad hoc decision making to rational business practices and policies you can use as you grow.

Who Should Write the Ad Spec?

In many cases, the ideal ad spec for the publisher at large is written with competing priorities in mind that balance the wants and needs of the sales and marketing teams with those of the engineering and development groups, not to mention what the Ad Ops team can realistically manage to support at scale.  The Ad Ops team should absolutely own the ad spec document, but it’s unwise to make ad spec decisions in a vacuum, without input and advisement from outside teams.  Just as it drives the Ad Ops team crazy when other departments make decisions that impact their world without having a seat at the table, Ops ignoring teams with a downstream interest in how the ads impact the site performance in general is a recipe for trouble.

The most productive approach is to define boundaries and a framework that the broad group of stakeholders can agree upon, and then let Ad Ops handle the due diligence on what ad technologies can reliably work within those limits.

Big Picture Stuff

From a technology side, I would submit you start the ad spec process by defining guidelines on latency and uptime – these metrics have the most noticeable impact on the user experience, and chances are, overall pageload time is what your web development and IT group care most about in terms of how ads affect their jobs.  Latency means how long it typically takes a 3^rd party technology to respond to a request, such as an ad request, and uptime means how often the technology will actually respond to request.  These are metrics typically outlined in service level agreements (SLAs) in technology vendor contracts, which Ad Ops won’t get from 3^rd party technologies, but can at least ask for as part of their due diligence in approving or certifying 3^rd party ad servers or rich media vendors.  For example, knowing if the 3^rd party technology has geo-located collocation facilities, meaning their servers are physically housed in diverse locations and at the high-speed fiber optic connections of an ISP, should separate the men from the boys so to speak.  Internal technology teams can not only help provide acceptable levels of performance, but can also weigh in with the right questions for Ops to ask in their evaluations.

In terms of sales and marketing, clearly outlining all the potential products, their configurations, and how they are represented in the marketplace should be the first step.  What ad sizes, formats, and functionalities will customers want and expect?  Going through this exercise can help compile an exhaustive list of products, which Ops can then use to define acceptable attributes like file size, expansion limits and directions for rich media units, and what types of creative is supported where, and for what platforms.  This is particularly important in today’s environment, when most digital publisher not only support desktop display advertising, but mobile, video, and tablet as well.  Perhaps mobile video units are only available for application takeovers, not mobile web, or maybe the file size limits for standard media are set at 40K, but interstitial units can go up to 100K.  Different rules for different products not only help the sales organization move signed contracts into campaign execution quickly, they can serve as QA cheat-sheets for the Ops organization in campaign setup.

In the next article, we’ll talk about how to move from general guidelines to more Ad Ops specific concerns, how to keep your ad spec current and flexible, and where you can view some of the most professional, comprehensive ad specs that the biggest digital companies use in their business.

In this post we are going to talk about ways to approach the “chicken and egg” problem of finding publishers and advertisers when you’re first getting started. Specifically recruiting publishers and managing the publisher agreements.

I am going to focus on the type of ad networks that I know best – traditional premium networks that work directly with publishers, serving either all or part of their traffic. These days with exchanges and DSPs you can start and run ad networks without ever talking to a single publisher – this post isn’t going to cover that scenario. (but perhaps a later one will)

Getting Started

You can’t go after advertisers until you have publishers committed to your network – advertisers aren’t going to give you the time of day if they don’t have a good idea of where their ads are going to be running.

In an ideal world you already run a successful, well known, and growing site in the space where you are starting an ad network. This is the ideal scenario because it would mean you have a working knowledge of the sites in your vertical and more importantly, they know who you are too.  If this is the case you can get started by simply selling direct advertising on your site to advertisers and then slowly adding in new sites that fit your vertical as you grow demand from your advertisers. In no time you’ll have a flourishing and growing ad network.

If you don’t already have a successful site with advertiser relationships how do you get publishers to join your network?

Commitment

One method is to collect commitments from publishers that will join your network once you have advertisers – this is pretty damn tough since most publishers hear from multiple people a month telling them they can make them more money and 90% of them never deliver. Publishers might also happily commit since it is zero risk to them – then never deliver when you actually sell some campaigns. I look at this as the method of last resort – there are better ways.

Single Spot

One of the most popular methods is to ask the publisher to give you a single spot on their site, 100% devoted to your network. This is popular since it is fairly low risk for the publisher since they are probably just going to add that slot in addition to their current slots. The benefit is that you can easily point advertisers to where their ad will show up – you have legit publishers on your network – and advertisers should take you seriously. The downside is that if it is an extra spot thrown on their site the placement could be bad, it could be next to other ads, and overall could hurt the brand of your network. You will also need to decide what to do with impressions you haven’t sold – you could backfill to another network, serve ads for your network, or pass those impressions back to the publisher. This is the second best of your options.

Passbacks

The best option is to get publishers to put your ad tags on their page and then to let them setup a remainder campaign for any impressions that you don’t fill – when you get started this will be most of the impressions. This gives you access to 100% of the publishers impressions, and even more importantly it gives you first shot at the impressions. This method also makes it very low risk for the publisher – the worst case scenario is that the publisher continues to make what they made in the past. (shameless plug: with some ad servers you would also get additional latency from the pass back, adOS is fast enough that publishers won’t notice this)

Minimum CPMs or Payouts

Some publishers will push for a minimum payout or CPM from you as a network – this can be very effective but obviously has huge risks. You can sometimes work out great deals with publishers where you give them some stability in the amount they make and also give yourself more upside (for instance giving a publisher a fixed $2 CPM but selling it for up to $5-8 CPMs). This can be very effective once you have your network off the ground – when you are getting started it is very risky. If at all possible I would stay away from arrangements like these until you have good advertiser deal flow and are confident you can hit the minimums (otherwise you can be under water very very fast).

So we have covered the methods of adding a publisher – but what about actually getting in contact with the publisher and even starting the conversation?

Brand

The first step is to ensure that you have a web-site of some kind up and running – you need to appear as legitimate as possible. This could even be a simple landing page using a service like KickoffLabs that looks nice, has a good logo, and describes what your network is about. Pay a designer – the $1000 you spend is well worth the investment (most networks don’t need more than a single page site). A large part of being a network is developing a brand – so this isn’t where you want to cut corners and be cheap or wishy washy. You should have a strong statement about what your network represents. Here is what The Deck says in a giant font on their site:

“THE AD NETWORK
OF CREATIVE, WEB AND
DESIGN CULTURE”

You need an equally strong statement that removes any doubt with publishers why you are a good fit for them. They should be honored to be invited to your ad network – if you look even the tiniest bit suspect it will sabotage your efforts.

Approaching Publishers

So you know how you want to on-board publishers, you have a strong site and brand setup. Now you need to actually reach out. The best way is to get in through an introduction – hopefully you know someone who knows someone at the publisher, or you know another vendor that the publisher uses, etc. If you can’t get an introduction then your best bet is a simple email, with a link to your network site, and a realistic estimate of the CPM you feel you can get for them. A 25% conversion rate would be awesome here – so you have to keep at it and invite more publishers then you need. (down the road publishers will be flooding you to join once you are successful).

If you’re well known in the vertical that you’re building a network  then you shouldn’t have many problems attracting publishers. For instance, I knew plenty of publishers before I ever started a .NET developer focused ad network. However, when I went to start a network in another vertical I had more trouble. I partnered with someone who helped me on the introduction side and I gave them a cut of the network profit for a period of time. It was a good way to get the network started and worked out well for both of us.

Once you have a number of publishers joining the network you are ready to start selling to advertisers – which will be the focus of Building an Ad Network Pt. 3

How did you get your first publishers? Share your thoughts in the comments.