Building and managing a vertical network of sites can be a big and expensive undertaking. The site identification, recruitment, marketing, and technical support all rely heavily on precision coordination and the effective allocation of resources and technology.
Once your network is up and running, you have finally built an asset, which can provide not only an ongoing stream of revenue, but can become a true asset for your business. Networks have been valued in the 100’s of millions of dollars in the market. When the network is functional and as you sell more campaigns into the network, there are a few risks that need to be managed. Most of the risks have to do with protection of your asset and publisher relationships while also ensuring your advertisers are kept happy and keep spending money with you.
Though many people in the industry focus on the risks of efficient inventory allocation and yield maximization, this is not the primary tactical risk for your network. If you run out of inventory when running a vertical ad network, you can go get more. Yes, it might take some time to find the right sites, but they are out there and there is no reason to focus so much on the inventory risk. The real risk is what is out of your control, which in the network world is also known as ad serving discrepancies.
One of the biggest issues associated with online advertising are discrepancies between advertiser impressions ordered and publisher impressions delivered. To exacerbate this issue, the discrepancies on a network are a tangible risk to your bottom-line and your operating profit. Most ad operations teams focus on discrepancies as an after campaign analysis. The standard approach is to do the standard quality control checks and perform some ad-hoc reviews of advertiser and publisher ad numbers during the campaign. The majority of any investigation or analysis typically happens at the end of the month or end of the campaign.
When running a network, discrepancy management and analysis actually has to be not only a part of the ongoing ad operations day to day, but also a big part of the initial tracking support. Because, in most cases, you will be paying your publishers based on how much your ad server/ad network platform reports, you can be at risk because every point of discrepancy in the numbers can have an exponential effect to your profits.
For example, say you had a campaign that was sold for $100,000 at an eCPM of $10 to Paramount Pictures for the 3 weeks prior to the release of a new film. You traffic the campaign in your network platform system at the full budget and at an eCPM of $7 (which gives you a profit of $30K as the network manager). The campaign runs over the 3 week period and you spend the full budget of $100K.
The next week you do your typical after campaign audit to match your ad network serving to your 3rd party or advertiser numbers. Though you had checked on the first day and had identified a 2% discrepancy, the gap had broadened due to the Rich Media unit being displayed having such high latency. You run a final report and your discrepancies have reached almost 10%. You only really delivered around $90,000 worth of the campaign. Because the campaign is perishable (due to being a movie release), all you can do is only bill the advertiser the $90,000 for what they say was delivered.
Now you have to pay your publishers, ad server, employees, sales commission and yourself. If you pay the reported number you will end up with a profit of around two thirds of your initial profit ($20K). Though the discrepancy was 10% (still a high number), your loss on the campaign was over 30%. When you take into account your commission payments, ad server costs, and other operational costs, this 10% discrepancy can actually decrease your operating margin by 60-70%.
Within the network world, there are three ways to manage this discrepancy situation. Well run networks will use a mixture of all three, but will really focus upfront on upfront discrepancy diminution, (UDD) which is the management of all discrepancies as a part of any campaign that is run on the network. The better you are at UDD the more you will be able to ensure publisher happiness and maximize your operating margins.
The three most common and acceptable methods to handle network based discrepancies include: 1. UDD – Upfront Discrepancy Diminution; 2. Ongoing Discrepancy Control; 3. Post Campaign Discrepancy Adjustments. Each of these methods are critical to the success of any network, though the networks that can manage the majority of their discrepancies through UDD will not only reduce ongoing campaign management and monitoring costs, but will solidify the trust relationship between themselves and the network members.
In the vertical network world the most common goal of a gold standard campaign would include the utilization of UDD. UDD is really the ongoing monitoring of discrepancies by ad server and creative type. The more robust you’re monitoring and tracking is the better you will be at implementation and execution of UDD. UDD is based on the historical tracking on your network. The way UDD works is through using an ongoing tracking database (could be as simple as a spreadsheet - download an example UDD spreadsheet here), and for each major type of ad unit keeping track of the stats around discrepancies on campaigns.
The major variables that you will probably want to track include: ad size, creative type, ad server, and if there was any targeting. You can make this tracking as robust as you want or as simple as you want. Many networks will just create one UDD factor and apply based on all campaigns.
The goal of this UDD analysis and tracking is to have upfront factors for trafficking. You want to identify the most likely discrepancy and manage against this number. The goal would be to reduce the need for ongoing maintenance to the campaigns for this purpose or even post campaign adjustments.
Once you have UDD tracking data, you can then use these factors in the actual trafficking. For example, say your UDD score for rich media campaigns was 8%. What you would need to do is traffic the buy in your network platform system at an 8% upfront discount. Publishers would never see this and would only be exposed to the slightly discounted CPM. A campaign that was a $10 buy would be discounted to $9.20 for the purposes of trafficking. If you think about what you are doing here, you are in essence adjusting for the discrepancy before it happens. What is the risk associated with doing it this way? Really the only risk is you still are estimating so there is a slight risk that the number would still be higher or lower.
In our previous example, you would have taken the $100,000 buy at $10 and discounted the CPM to $9.20. Now you have a buy that is $92,000 but will still deliver the same number of impressions and you use the same share of the buy on the eCPM basis. At the end of the month you have the 10% discrepancy we described earlier. You are still slightly off, as in this campaign you under estimated the discrepancy, but your margin has been protected.
In the first scenario, without the utilization of UDD, your margins were reduced by over 30% and your potential profit was reduced by almost 70%. Because you were proactive in the trafficking and utilization of UDD, now your overall hit was less than 7% and your operating margin was only hit by a few more point. Controlling discrepancies is really a function of protecting your business.
If UDD doesn’t fully cover problems or if you are still trying to find the correct UDD settings, there are a couple of other approaches to solve for discrepancies. The second in the list is really around the ongoing monitoring of discrepancies. Once a campaign goes live, your team needs to make sure they are actively reviewing the ad network numbers against the advertiser or 3rd party numbers. With active review you can continue to tweak and modify the UDD factor in campaign, and ensure you are protecting yourself and your publishers.
Finally, if something goes wrong, or if you just have not been able to utilize UDD or ongoing monitoring, the last option is to do a post campaign adjustment. Most publishers know that the ‘live’ numbers provided to them during a campaign are estimates. Savvy publishers also are very aware of all the variables that can cause numbers to be incorrect. At the end of the month or campaign you can apply a post campaign adjustment and communicate this to your publishers. This approach is not considered the best and should be avoided at all costs, but there will be times where a post campaign adjustment is required.
Publishers have said many times that they would much rather have a slightly lower CPM than to have the numbers at the end of the month change. When your network members login they want to rely on the data and trust it. These sites might be smaller, relative to your sites, but they many times are independent publishers that are small businesses and are relying on this revenue stream to manage their business.
Part of running a network is risk. The reason many networks retain upwards of 50-60% of the ad revenue is because they carry the majority of the risk. That being said, the way to ensure a high quality, trusted network, that also has a great growth rate and operating profit, is through proactive management of this risk.
The famous quote by MacArthur ‘There is no security on this earth. Only opportunity.’ says it all!
success by lift in metrics like awareness, message association, brand favorability, purchase intent), finding blogs that have real equity with their readerships is first and foremost. 
