What is Ad server and what is Adnetwork?

Ad server is a tool or interface used by Ad Agencies and Clients to make easier ad trafficking, reporting, optimization and ad performance. Ad server generally provides different data like Impression, clicks, Click Through Rate, Leads, downloads, video tracking etc.

Ad network is a company which connects advertisers to web sites that want to host advertisement, also this is a single buying opportunity for media buyers, an opportunity to access advertising market for small publishers. Ad network should have their own Ad server

DCM: Custom Floodlight Variables

Custom Floodlight variable is an advanced feature in Floodlight tag by using this we can capture the information beyond the basic (Visits and Revenue). Now a days you have the scope of setting 100 Custom variables per Floodlight configuration.

Adding the custom variable tags would involve basically the floodlight with an extra parameter in the tag which you would need to have filled in dynamically as the page loads, depending on the option a particular user is looking at. With this set up, the total conversion data would be available in the Custom Floodlight Variable report for the breakdown.

Below are the steps to enable a custom variable for a Floodlight:

1) Select the 'Floodlight Configuration' tab for an advertiser
2)Scroll the the bottom and enter a friendly name for a custom variable
3) Open the Floodlight under the 'Floodlight Activities' tab and select
the required u-value under 'Include Custom Floodlight Variables'
4) You should then see something like 'u1=[Subs Type]' added to the
Floodlight tag in the preview.

The webmaster is responsible for dynamically inserting the correct values
into the Custom Variables.

You should keep in mind the following points:

1) DCM is unable to read and process some special characters and if you try to pass special characters as part of the custom variable values, the populated values sometimes do not show up in reports, or they show up incorrectly.

2) Avoid passing special characters into custom variables. However, if youcneed to use a special character as a separator for multiple values, use acpipe (|). For example, if you try to pass product names to a customcvariable by replacing u5=[productnames] withcu5=product1+product2+product4, you'll produce an error. Instead, the valuecshould be passed as u5=product1|product2|product4.

Please note, that if you add the variables to existing Floodlight tags, you will not lose any previous data, the tag will simply start including the new variables (and their values) in the Custom Floodlight Variable report. After adding a variable, remember to resend the Floodlight tags to the webmaster so that he implements the modified tag on the site.

Why is it not appropriate to compare Google Analytics and DoubleClick Campaign Manager data

It's important to note that conversion rates for each product will differsince each product uses different measurement systems. Since DCM and Google Anaytics are two different products and they are not integrated, the numbers are expected to be different.

I understand not being able to compare the data sets is frustrating, but I can tell you that better integration between the two products is on the way.

In Google Analytics, the Conversion Rate represents the percentage of visitors that convert on at least one of the goals defined for a particular profile. Only one conversion is counted per visitor session,
even if they convert on several goals, or one goal more than once. DCM records conversions based on the type of activity you had selected (standard, unique, sessions). Also, Google Analytics URL tagging method records only post click data.

Here are some other differences between the two products:

- GA uses Referring URLs which not a reliable way of determining DoubleClick-driven traffic

- GA requires acceptance of JavaScript, DCM does not

- GA requires cookies, DCM does not. DCM can use IP addresses when cookies are not available.

- If a user clicks on an ad twice within thirty minutes without closing his or her browser, this will be registered by Analytics as one visit to the site; whereas DCM would record two conversions.

- GA does not filter for fraudulent data

Also note that DCM applies algorithms to identify robot and spider clicks.

Definitions:

Visits - represents the number of individual sessions initiated by all the visitors to your site. If a user clicks on your ad twice within thirty minutes without closing his or her browser, this will be registered by Analytics as one visit to your site, even if the user left your site and then returned shortly after.

Visitors - represents the number of unique users that visit the site on a daily basis. Any sessions from the same user on the same day will be aggregated into a single visitor, but may represent two or more separate visits.

Page Views - defined as a view of a page on your site that is being tracked by the Analytics tracking code. If a visitor hits reload after reaching the page, this will be counted as an additional page view.

DCM: Click to Landing Page Discrepancy - Goes up

The reason you see a higher number of post-click activities than clicks recorded is due to the fact that the counter type of the floodlight tag is set to 'Standard'. This means that DCM will count each visit to this webpage as a separate post-click activity.

Let's imagine a user clicks on the ad and goes to the webpage, we have 1 click and 1 post-click activity. Now let's say this same user goes back to the landing page but through a different medium rather than a DCM ad (through Google, or a bookmark, etc), we will record another post-click activity. This means we will now have 1 click but 2 activities. This is how you can see more post-click activities than clicks.

You can change the counter type to 'Unique' which means that DCM will count only the number of conversions by unique users within a 24-hour period. This means that if a user visits the page multiple times within 24 hours, we will only count one post-click activity. If, however the user visits the webpage again, outside the 24 hours, we will record this as a separate activity.

If you want to ensure that only one post-click activity per browser session is counted, then the advertiser is responsible for dynamically inserting a session ID, typically based on data included in a session cookie served to the user by the advertiser. You can set the tag type to Per Session but you will need to speak with the webmaster about whether this is possible as they will need to implement this at their end.

After making changes in DCM, please remember to send the newly generated tags to the site.

DCM: Marin Integration and Data transfer

Marin integration can be done by using piggybacking dynamic pixels with DCM floodlights, and in this case, Marin records conversion data (including revenue) directly using their piggybacked pixel.

If the integration is done in this method, I’ll also need to corresponding floodlight IDs which has the piggybacked Marin pixels. I’ll pull up a report for these activities for the dates in question, and verify if there were any conversions for those days.

Alternatively, Marin could also use DCM Data transfer files, and extracting data using their own system. Just to give you an overview, DCM generates server logs for the data transfer enabled network, and generate 3 log files daily. One for impressions, one for clicks and one for Floodlights(conversions). These contain data for the all the advertisers within the network. As these logs are generated daily, if there’s no impression/click/conversion activity for a specific advertiser on any given day, its ID won’t populate in the log file.

There are referenced using match tables, and the required data is extracted/attributed using separate tools. Just to let you know, the naming convention for the data transfer log files is below:

NetworkImpression_[network_ID]_MM-DD-YYYY.log.gz

NetworkClick_[network_ID]_MM-DD-YYYY.log.gz

NetworkActivity_[network_ID]_[Floodlight_ID]_MM-DD-YYYY.log.gz

If you could let me know the name of the log file that Marin is using for any of the date, I’ll be able to lookup using the same in your Google cloud storage (GCS) bucket, and confirm availability of the log files. Based on the above naming convention, your GCS bucket has all the three log files added for the dates in question. I've attached two screenshots for your reference.

It’ll be helpful to know how Marin is setup, in order for me to pinpoint the issue. Please let me know if it’s possible to discuss this over a call, and if so, the time that works best.

DCM: Video completes are exceeding video starts, and impression counts.

The below response could be useful when we have any issue wherein the video completes are exceeding video starts, and impression counts.

I understand that, we’re seeing discrepancies for in-stream Placements, I’d like to share that, the root cause of the issue here is related to our enhanced traffic quality filters. This is a known issue, as we aren’t filtering the enhanced video events, however, we’re filtering impressions (I’ve seen similar issues reported to us earlier), that’s why the video completes are high than the impressions.

DoubleClick has developed a sophisticated set of algorithms that determine which clicks/impressions are invalid. According to an analysis, there’re three major reasons for these invalid events:

Spiders, robots, crawlers and other automated agents
Programs that download websites for offline use
Link analyzers and code validators

Spiders, crawlers and other automated programs are often, though not always, designed to first check for a file called robots.txt before sending a request to any server. This file sends instructions to the automated program that can tell the program not to make any requests to that server.

DoubleClick uses robots.txt files to help prevent clicks from automated programs. This file doesn’t, however, prevent downloading programs and link analyzers from generating automated clicks. If DC receives repeated, successive clicks/impressions from the same user within a set time frame, then it concludes that the user is an automated agent and not an actual person. These clicks/impressions are determined to be invalid, and thereby designated to be discarded.

An analysis has shown that very few impressions by real users are discarded using this algorithm. Once a user is identified as an automated agent and the impressions are invalidated, all subsequent clicks/impressions that occur are also discarded.

Further, to give you more details, our number one priority was to implement enhanced spam filtering for impression and click data then the enhanced metrics.  Our internal teams are aware of the discrepancies, and working to resolve them over time. That being said, we're working towards resolving these discrepancies in Q2, though, we don't have a more official timeline. Once these filters are in place, I'm pretty sure, there won't be any mix up

Floodlight tag can't be implemented in email

If you are trying to implement the Floodlight tag in email body. Please read this article completely before proceed.

I'm afraid that won't work in most cases. The mail stage is an exposure, not a conversion. The Floodlights are recommended to stay on the site because a "conversion" by definition is an action on the site or a buy.

Additionally, the email campaigns are never tracked 100% correctly. Please see the explanation below:

- IMPRESSIONS ON 1x1:
Only a fraction will be recorded since in many mail systems the images are not displayed by default. When called the tags, they can't have a random number anyway (no JS), so you will undercount in any case. For both these reasons, a massive underestimation of the opened mails will be produced.

- CLICKS:
If the user clicks on links tracked by click trackers, this action will be indeed counted. However, it is not guaranteed that the following conversions will be recorded, since the user can click from an email client (Outlook, Thunderbird etc.) and open a default browser (IE, Firefox etc.) which could have a different cookie set.

Apart from this, implementing Floodlight tag into the email body is not recommended because:

1) Only Image Floodlights will have a chance to be fired.
2) No possibility to generate random numbers (under counting).
3) By default, many clients are blocking images anyway.