Category Archives: Publishers

Mar 03 2012
Leave a comment
Ad Ecosystem, Ad Serving, Ad Technology, Publishers

Advanced Ad-Serving Features, Part 1: Site-Side Servers

(Originally published in ClickZ, November 2001) by Eric Picard

A few things have happened since my last article:

  • DoubleClick did not buy Real Media, as was widely speculated.
  • 24/7 Media did buy Real Media.

With all the turmoil in the ad tech market, it might be time to review your tech partners’ stability. You can read my comprehensive (if a little dry) recommendations from way back in June. Boy, it sucks being right sometimes.

Last time, we talked about the basics of ad serving. I got a few emails from ad-serving companies arguing that I didn’t cover enough in that article. That was the point, actually, to write a quick overview. I doubt I’ll hear that complaint again — this two-part series goes into plenty of detail.

I will say up front that this is still a generalized overview. There are individual features of various ad-serving products that I won’t be covering. You still need to do a comprehensive review of various offerings before making a final technology decision.

I did get emails from individuals who couldn’t see much specific differentiation between a site-side and a third-party server. One wrote: After reading your article, I still do not see the advantage of having a third-party, other than the same reports in the same format… What more do third-party servers provide other than the number of impressions delivered and the number of clicks? Is there some other type of analysis of the campaign that the third-party provides?

First, I am not advocating the use of third-party servers over site-side servers. The two types of ad servers are designed with different purposes in mind. Site-side and third-party servers are not competitive. Site-side servers are aligned to publishers, while third-party servers are aligned to advertisers and agencies.

This analogy won’t earn me any points from the site-side server companies, but we could put it like this: A site-side server is to a third-party server as a freight train is to a passenger train. Both must be able to travel on the same tracks. Both must travel at the same speed. Both must deliver their content accurately and on time. But the passenger train needs to be a bit more refined in its amenities. The freight train needs to be able to handle a heavier load and deal with different delivery protocols (after all, passengers walk off their trains while cargo needs to be unloaded).

Second, unified reporting and trafficking procedures may not seem like a big benefit if you’re an advertiser that doesn’t do lots of trafficking and reporting for large media buys. But if you have to integrate 20 or more unique site reports into one single report for a client, it isn’t a simple process. It can take days or even weeks.

Otherwise, the reader is on the right track (pardon the railroad reference). The real strength of a good third-party server comes from its advanced features. There are plenty of advanced features to discuss on the site-side as well.

So let’s go deeper. This time, we’ll look at site-side servers, and we’ll go into depth about third-party servers in my next article.

Site-Side Ad Servers: Advanced Features

Just to reiterate: Publishers use site-side servers (sometimes called local- or sell-side servers) as part of their business infrastructure to accurately deliver and report on ad delivery. This includes trafficking controls, workflow, inventory management, and many other parity-level features. Many (but not all) site-side servers include the following features.

Targeting

  • Geographic targeting. This feature works best when applied on the site side. The ad will be sent to only those users who are where you would like the ad to be seen. The main issue in geotargeting is accuracy. There is wide disparity in the accuracy of methods used to target based on user location. Not all solutions are created equal. Some subitems include targeting based on IP address, Standard Industrial Classification (SIC) codes, and ZIP Codes. 
  • Profile-based targeting. A few of the bigger players spent millions trying to build accurate databases of user activity so that ads could be targeted based on Web-surfing habits. You can, for example, send an ad for a motorcycle to people who visit sites for motorcycle enthusiasts. This is sound in theory. In practice, it’s difficult to implement and very expensive to maintain — a critical consideration, given current market conditions. Ask how frequently a vendor’s database is refreshed and what the average age of its profiles is. Some experts question the validity and value of profile-based targeting, while others claim success. 
  • Content-based targeting. This is a feature generally offered by portals, networks, and search engines. Subcategories might include keyword search results and content categories. 
  • Session-specific targeting. This includes domain, browser-type, and operating system.

There are other types of targeting that you may come across, but the above are the primary methods.

Creative Rotation Controls

  • Frequency capping of creative. This allows you to specify how many times you want an individual user to see creative before you “shut it off.” Experts recommend frequency capping be used to limit the number of exposures an individual user will have to your campaign. Being able to cap frequency is the culmination of the desires of brand advertisers from offline media, which they cannot do offline. Direct marketers should take note, because it is possible today to review the effect of frequency on conversion. 
  • Sequential serving of creative. This feature lets you specify a sequence of creative elements shown as the users travels across Web pages. For example: A car drives along a highway with a tantalizing opening message. Next, the same car appears with the second part of the message. Finally, the last message appears with a hook or call to action to draw users in. 
  • Accounting interfaces. Some site-side servers include interfaces for popular business accounting packages, such as Microsoft Solomon.

You’re on your way to becoming an expert. Next time, we’ll examine the even-more-advanced advanced features of third-party servers.

Mar 03 2012
Leave a comment
Ad Ecosystem, Ad Serving, Ad Technology, Media Agencies, Publishers, Rich Media, Startups

Protecting yourself from exploding Ad Technology partners

(Originally published in ClickZ, June 2001) by Eric Picard

As you might have heard, AdForce closed its doors last week. CMGI shut it down after failing to find a buyer. What a strange world we’re living in. It makes you wonder whose balloon is going to pop next, doesn’t it? It wasn’t so long ago that these companies were awe inspiring to many of us in the industry. How quickly things change.

So what is an agency or advertiser to do? If you make decisions about or manage the relationships with your technology partners, maybe we can come up with some guidelines for you on how to judge a company’s stability. The world is a dangerous place right now, and getting caught unprepared if your ad-serving infrastructure suddenly goes up in smoke could really hurt.

Public Companies

Let’s start by looking at the public companies. Once, you could assume that if a company was public, it would most likely be around in a few months. Not any more. Being public actually works against most companies in our space these days.

A very well-known industry analyst once told me that any company with a stock price of under five dollars is focused only on appeasing the investors, not on doing business, and certainly not on innovating. At the time, he said he didn’t even pay attention to those companies. I have to wonder if he still operates on that principle. Today, companies in our space with stock near or above five dollars are like superheroes.

I just looked at the list of 43 stocks I’ve been tracking for the past three years. It’s made up of companies either active in online advertising or peripheral to it. I noticed that 34 are trading under five dollars or are no longer trading at all. Of those 34, 12 are trading under one dollar, and 17 of them are no longer actively trading. Some have stopped trading because of merger or acquisition, but most are just out of business.

So how do you make sure you’re safe?

  • Look at the makeup of your technology partner’s customers. And I don’t mean its “portfolio” — since this often contains customers no longer working with the company. I mean active customers.
  • Ask your contacts for active customer references. This might be a tough play, but at least you’ll be able to judge the stability of the company. If a company doesn’t have one customer willing to say something good about it, you might want to reconsider.
  • Find out how much operating capital the company has. DoubleClick is an example of a public company in our space with lots of operating capital. 24/7 Media is an example of one without any operating capital.
  • Make use of the fact that the company is public. Look closely at its public disclosures. Read quarterly reports. Read analyst reports. Judging which companies are in trouble is mostly a clear and commonsensical act.
  • Check if the company traded as an over-the-counter or bulletin board (OTC:BB) stock. If so, you should be especially wary, because the regulations about reporting here are much less clear. Since there is less regulation, companies trading as OTC:BB are often seen as stock scams at worst and as a little shady at best.

    The main change these days is that since many companies have been de-listed from the Nasdaq, as victims of the times, they have ended up on the OTC:BB. If the company is a real business, you should be able to tell pretty easily. One quick test is to verify that its “gallery” or case studies are real, not mocked up. Be direct and ask — even ask for real customers you can talk to if you feel uneasy.

    One of the first questions I ask a company is if it is public or private. If it’s a small company and it’s public, I immediately ask if it’s an OTC:BB.

Private Companies

Now that we’ve looked at public companies, we should review how to judge the stability of a private company. It’s not so different, but some of the information isn’t available publicly.

First, make sure that the general items are covered from above — and especially focus on customers. Since public companies are accountable for things that they say in public, they usually are relatively credible (minus the marketing spin). Private companies are not so tightly regulated, so make sure to do your due diligence.

Usually, private companies’ financial health is the hardest point to establish. And today, this is the most critical factor to review. There are, usually, some indicators:

  • Private companies generally start up through bootstrapping or venture financing. If it’s the latter, you’re in luck. It’s a huge win to get investment from a venture capital (VC) firm, and the general response is to issue a press release.
  • Review the company’s press releases, and try to figure out how much money it’s raised.
  • If it has been growing and hasn’t raised any money in the past six to eight months, your warning bells should go off. The only situation in which that shouldn’t worry you is if the company is bringing in lots of revenue. This is tough, given the market right now.
  • Next, you should figure out how much time it has to get profitable. Once you do (explained below), feel free to ask the company directly how it plans to achieve profitability. Again, you may not get an answer, but it doesn’t hurt to ask.
  • Get a general idea of the company’s burn rate by using the following guidelines:
    • How many employees does it have? Usually, the company will tell you.
    • Where does the company have offices, and how many people are in each?
    • What kind of capital expenditures might it have? If it’s an ad-serving company, how much is it spending on server farms?
  • This is closely held competitive data, so a company is unlikely to just hand it over, but you can make some educated guesses about these things:
    • Figure that a company with 30 employees is burning $150,000 a month for salaries (if it’s aggressive).
    • If the company is in New York, figure that it’s paying big-time rent (even with some of the new deals opening up). An office for 30 people will run roughly $35,000-45,000 a month, depending on location and other costs beyond rent.
    • That gives us close to $200,000 monthly without even getting to hardware or server infrastructure.
  • When you look at (a minimal) server infrastructure and costs for setting up and maintaining the business, we’re talking about $250,000 monthly, or about $3 million a year.
  • That gives a company of 30 about two years of life if it’s raised $6 million. You can work out the various scenarios for different sizes and funding.
  • I’m being quite conservative here, and this is based on a whole lot of assumptions. For instance, if the company happens to have offices in New York and San Francisco, you can imagine that the costs are a lot higher.
  • If the company is too large for the amount of funding it has, it will burn out fast. If it’s too small for the amount of business it has, you’re going to get horrible service.

So now you’re an expert in evaluating those companies you’re working with. Go out and look at them. There’s no time like the present.

Mar 03 2012
Leave a comment
Ad Ecosystem, Media Agencies, Publishers, Rich Media

Analytics: Beating Ad Clutter

(Originally published on ClickZ, March 2001)

The outlook for online advertising is good; we just need to fix the problems. A typical Web site publisher runs an ad on every single page of the site. If it has unsold inventory, it either runs a public service announcement or it runs a house ad.

Jupiter Media Metrix just released the report “Online Advertising, Market Perception vs. Reality.” It was very enlightening.

While growth has slowed since online advertising’s meteoric rise over the past few years, the industry is still developing at a healthy clip. And as I’ve been saying since first noticing this dip, it’s not like people have stopped going online: There was a 22 percent increase in the number of unique users online between 1999 and 2000. More important, the amount of time users spend online went up 31 percent in that same time period.

With the consolidation of sites as dot-coms fail and the resultant concentration of user traffic, we’re going to see more value placed on those sites that survive this dip. This is very good news for the long term.

Information Overload

One point made by Marc Ryan, director of media research at AdRelevance, really caught my attention. He said, “Publishers should run house ads more sparingly, surgically, to reduce clutter and allow clients’ ads to have more impact.”

Publishers are selling less inventory, and the inventory devoted to house ads on their sites is increasing. One of the greatest problems online ads face is user information overload. Ads appear over and over in a site’s design in exactly the same static ad space. The theory is: If we leave it empty part of the time, it will draw more value out of those ads that do appear.

This notion hit me pretty hard; it’s basic stuff, but I hadn’t thought about it this way. And as is my wont, I started thinking about the problem from a technology standpoint.

The first step is to validate this theory — it sounds logical, but so have lots of things that haven’t been proven to work. So how would we validate our theory? If publishers can figure out what the effect of NOT running house ads is, then they can make intelligent decisions about best practices. But…

Optimizing Inventory

The majority of publishers and even agencies haven’t invested in analytics solutions — and I’m not talking about a roomful of people looking at campaign stats. That kind of analysis isn’t helpful without a good tool to sort the vast amounts of data before the human eyes start looking.

Part of the problem is that these solutions are extremely expensive and require a staff of people who are experienced analytics specialists. Virtually no agencies have this kind of staff. And the scary part is that the vast majority of publishers don’t have this kind of staff, either.

Yet analytical tools would solve many problems for publishers. They would be able to quickly and easily see value across their properties — and they could actually optimize their inventory differently. They could build new pricing and inventory plans. Imagine what would happen if publishers could show a value proposition to media buyers that leads them through their buys, AND the end result shows a greater ROI.

For example: Traditional media planners are used to buying broadcast media in time slices for specific channels, whether it’s a radio station or a television station. For the most part, we haven’t done this online, partly because of technical reasons and partly because of the absence of a value proposition showing that it works. If the publishers could make this work technically and run some tests, they could analyze the results and see if there’s a business case to be made.

Without analytics capabilities, publishers can’t possibly know what works and what doesn’t. Essentially, most have been flying blind.

So what’s the answer? Publishers need to dive deep into analytics over the next year. It isn’t a critical issue yet — but expect it to start looming soon. And they need to start doing research into what kind of questions they need answers to.

My advice: Please don’t make the mistake of thinking it’s about click rate. And do include rich media ads in the mix when doing the analyses. This is another reason tools are needed, because if you can’t figure out what’s happening with GIFs, how are you going to figure it out for rich media, with exponentially more data to crunch through?

While the industry is not doing nearly as poorly as some fear, it’s time to start cutting through theGordian Knot to solve some of the problems that clearly do exist.

Newer posts →