Demand Side Platforms and which one you should bet on
If you’ve been using ad platforms like Google Adwords, Facebook, Linkedin, Twitter and so on for a while and are wondering how to access additional inventory or get greater control over your campaign, you are looking for Demand Side Platforms.
The world of programmatic advertising is made up of Demand Side Platforms(DSPs) and Supply Side Platforms(SSPs). Demand Side Platforms are what advertisers use to bid on ad inventory while Supply Side Platforms collate ad inventory from publishers. The transaction takes place on an exchange through a programmatic instant auction through Real Time Bidding. Simply put, when an impression is generated, advertisers compete for it through Demand Side Platforms and the ad of the winning bid gets displayed. You can imagine how quick the entire process needs to be to seem instantaneous at the website/app visitor’s end.
If your digital marketing campaign involves Google Adwords, Facebook, Linkedin, Twitter ads as well as some direct buy tie-ups with publishers in your geography, you’re already targeting anywhere between 30-90% of the inventory available online in your geography. However, using a DSP has 4 key benefits:
- Access inventory from all Supply Side Platforms
- Serve direct buys from publishers through the platform for greater control and targeting
- More detailed targeting like which fold of the page the ad inventory appears on
- Many more optimization algorithms available
I evaluate Demand Side Platforms on the following criteria:
1. Ease of use
A DSP is a much more complicated platform to use in comparison to self serve platforms like Adwords or Facebook Ads Manager. If you don’t have a marketing team who has worked on a DSP before, it is going to be a slow learning curve. Creating and Managing campaigns on a DSP also are more time consuming and you need someone who is really good with data to make sense of the mountains of data you will get from your DSP reports. While ease of use may not seem like critical to the performance of the campaigns, I believe it dictates how well your team will end up using the DSP and hence will impact the performance of the campaign eventually.
2. Optimization algorithm
DSPs use a lot of machine learning and automation to create algorithms that can help an advertiser achieve his goal. The efficiency of these algorithms thus has a direct impact on the performance of an online ad campaign. Some DSPs give a lot of different optimization algorithms while others give a few. An advertiser needs to test a DSP’s optimization algorithm because this decides whether ads are being shown to the right people or not.
3. Targeting options
As the programmatic world is moving to 1 on 1 marketing, DSPs are tying up with data providers for their advertisers to target. DSPs also allow for different targeting options such as the minimum % of users who are likely to see that ad inventory item. As these targeting options differ from DSP to DSP, it is important to understand how narrowly you can target your consumer through this DSP.
4. Inventory available
A lot of advertisers complain about the amount of fraud inventory available online. DSPs are working with third party authentication services to ensure that junk inventory is not served to their advertisers. It is important to check with your DSP about the steps they’ve put into place for this and the remedies available if fraud inventory is served to the advertiser. Additionally, another question worth asking is apart from the open exchanges, which supply side systems does your DSP have access to because this can give you access to additional inventory.
Getting a campaign to be successful depends a lot upon the support you will get from the platform provider. If you do not get high quality resources for the first few months to help set up the campaigns, you may not get the results you invested the money for.
Here’s a quick snapshot of Forrester’s analysis of DSPs:
You can download a copy of the report from Dataxu’s website through this link. Dataxu and Mediamath both have excellent platforms. While this quarter may put Mediamath slightly ahead, Dataxu has been ahead of Mediamath for a few quarters in the past as well. Both Dataxu and Mediamath have very easy to use platforms with a good UI. Mediamath is probably better in terms of look and feel but Dataxu offers more in terms of optimization.
You may also be tempted to get Google’s DoubleClick offering. Our experience with DoubleClick has not been very satisfactory. Their platform is extremely complex to learn, reporting is very slow, optimization is not that great and the support we got didn’t help us much.