Prospecting and Remarketing in programmatic: How to make use of them [Part 2]

This is the sequel to part 1. Make sure you start with part 1 if you haven’t read it.

If part 1 of the series, we covered off the on the 2 basic types of targeting settings in a programmatic campaign –  prospecting and remarketing. We then covered off 4 targeting settings within prospecting.

In this part, we will cover off some more prospecting targeting settings.

Viewability targeting: 

Viewability targeting allows you to target your ads in such a way that allows them to be in view (on the screen) of the user when served. The MRC definition of viewability is 50% of the ad in view for a minimum of 1 second for display banners and 2 seconds for video (more on video later). This means that at least 50% of a display banner needs to be in-view on a user’s screen for at least a second for the impression to be considered view-able.

If you think that 50% of the banner on screen for at least a second is way too little for that banner to get any attention, some viewability vendors allow you to customize what you consider to be a viewable impression in both time in view requirements and % of banner in view.

How does viewability help your campaign’s performance? It allows you to make sure you are targeting (and depending on how you negotiate, paying) for impressions that are actually seen, rather than targeting impressions on ads which are below the fold, hidden away from the sight of users.

However, does viewability = better performance? In a perfect world, yes viewability should equal better performance. After all, how can a banner that has never been seen drive any performance at all?

But in the real not so perfect world, a bit of math is in order to figure out if your viewable inventory is driving better performance.

Let’s run some hypothetical numbers!

site x is selling you impressions at 1$ cpm

50% of those impressions are viewable on average.

You go to site x and request to buy 100% viewable impressions from now on. Site x agrees and offers you a 2$ cpm for 100% viewable inventory. The publisher still has to pay the same bills after all, but at least you have a guarantee regarding the quality of inventory you’re getting from now on. Great!

Will your performance improve?

At first, site x CPM was 1$ with 50% non-viewable. That’s 2$ cost per viewable impressions. Basically the CPM for viewable impressions stayed the same. Fair.

But now, there’s your tech cost to measure whether the publisher is upholding the 100% viewable impressions agreement. Say that’s 0.1$ CPM paid to a viewability vendor to monitor this. So you end up 2.1$ for viewable impressions without caculating the man-hours running reports to make sure you are actually getting 100% viewable impressions. This is a marginally worse deal (and in some cases I’ve seen deals that are way worse)

This is an all too recurrent theme that I’ve seen. Am I saying do not use viewability tools? No, by all means you should use them. But when you do, run the numbers. Make sure you’re using those tools to improve the deal you’re getting rather than ending up with a bigger bill.

A word on viewability in video (and video targeting in general): Video has challenges in viewability due to the VAST(video as serving format) standard. This standard does not support viewability unfortunately and most available video inventory is VAST at the moment (YouTube being the biggest culprit). As a result, viewability reporting on video isn’t great.  Video has some other targeting features, such as player size, position, etc. I’ll cover those a bit later in a video specific post.

3rd Party data

3rd party data is data that the advertiser (say Mom’s Pizza and Hardware) does not own, but instead buys data from a data vendor. This data is normally in the form of audience segments with demographic, interest, or contextual data attached.

An example of a segment could be: “Women”. Another example could be “age 18-24”. Another example could be “Loan prospect”. There are 3rd party segments about almost everything and they can be really hit or miss. What they have in common is that you pay for them. You could pay for these segments on a CPM basis or you can pay on an all you can eat basis (fixed fee per month and you use all the data you want). 3rd party data provider examples are: Bluekai, weborama, exelate, vdna.

In your campaign, you’ll want to break those data segments into separate line items. This is to be able to detect which segment is performing and which isn’t. That’s not to say you can’t combine segments (combining “women”  AND “age 18-24” segments will let you reach women aged 18-24) but always try to have a way to gauge segment performance on an individual level so that you can later optimize.

3rd party data has not often worked for me except in reach campaigns (where your goal is to reach a target audience). That’s because the cost of 3rd party data often offsets any efficiency you might expect to get out of it. An all you can eat approach to buying the data could be a way to mitigate this if you’ve got a big enough budget to spend. Each advertiser will find different results, so the only way to know for sure how it will perform is to test test test!

Day of the week and time targeting (aka dayparting)

Day of the week and time targeting are powerful tools that are often underused. Depending on your product, time of the day can have an immense effect on your campaign’s performance. The simplest example would be if you’re managing a campaign for a pizza delivery place – around lunch time and dinner time you’re going to see peak performance, with dips in between. Using time of day targeting, you could increase the delivery of your ads in the run-up hours to lunch and dinner, and perhaps exclude the hours of the day where your restaurant is closed.

Similarly, if you close on a certain day of the week, you can use day of the week targeting to automatically pause your ads on that day.

There’s also benefits to be gained here for advertisers that would not necessarily think day of the week or time of the day is important to them. The easiest way to see if you can get more efficiency out of your campaign using this targeting setting is to run reports broken out by time of the day/day of the week. You might be surprised at what you find – often there are dips in performance at certain hours on certain days that you will  be able to detect and use to adjust your bidding.


This brings us to the end of Part 2. Sign up to get a notification when Part 3 is out.

[mc4wp_form id=”520″]

3 thoughts on “Prospecting and Remarketing in programmatic: How to make use of them [Part 2]

  1. Liked how you give practical advice. From what i’ve noticed managing sales driven campaigns, it seems some advertisers are prioritizing cookie dropping and conversions, over weighing in the actual viewability as a metric. What’s your p.o.v. on that?

    1. Thanks Daniel! Yeah agreed it seems that advertisers often do not take viewability into account while setting up their attribution models. In my view a non-viewable impression should not be considered while looking at attribution. The challenge is getting reliable viewability measurement in that case and finding a satisfying threshold for viewability, but ideally if an impression was never viewed on screen it couldn’t have affected the conversions/sales imo.

      1. Agree and it’s not an easy discussion to start if that means we are potentially not counting at least half of the impressions. So far I haven’t seen any attribution model that fits the given description but I do believe it deserves more attention

Leave a Reply

Your email address will not be published. Required fields are marked *