Posts Tagged ‘DSP’

It’s powerful, flexible, customisable, saves thousands of man hours, provides valuable customer insights / behaviour and most importantly ensures that you get a healthy ROI if used in the right way.

Meet The Brain: The Brain is MediaMath’s proprietary algorithm and ingests data (60 billion opportunities everyday to be exact) and decisions against that data.

Their algorithm’s left-brain and right-brain work together to analyse large pools of impressions, looking at dozens of user and media variables, to determine which impressions will best meet an advertiser’s goal.The Brain values each impression based on its likelihood of driving an action, and bids accordingly.




When a process is automated and self-service is the definition of programmatic.

When looking at programmatic buying, this would include: paid search, Facebook marketplace and FBX, affiliates, display RTB, online video RTB and any other biddable media / RTB buy.
Trading Desk

A team of ‘programmatic planners / buyers’ is the definition of a trading desk.

Most agencies claim they have a ‘trading desk’ yet run paid search, affiliates and Facebook marketplace in different buildings to their RTB buys, in which case they should be called a managed service DSP or ad network.

Only when an agency has all programmatic buying under one team and one ‘head of’ can they truly call themselves a trading desk and have the ability to run a trading desk as it should be done.


Firstly you need to have access to a DSP and have adserver container tags across your whole site. When implementing the container tags, it’s essential to pass back as much data through the custom variables as possible eg. age, gender, bucket amount, revenue, customer loyality type.

Container tags should be placed across each site / product page and then a tag from homepage throughout the whole conversion process to the sales thank you page. Also a tag across the current customer login section is required.

Now it’s a case of building up your CRM database within the DSP. A pixel within the DSP represents a targeting segment to be included / excluded such as:

  • Main acquisition homepage
  • March 2013 microsite landing page
  • Car Insurance homepage
  • Home Insurance homepage
  • Quote confirmation page
  • Logged in
  • Deposit confirmation page
  • Business awards landing page
  • Affiliate landing page
  • CRM email – non converting
  • Males
  • Age 25-34
  • Gold customers

Once your pixels have been created in the DSP, it’s a case of implementing them within the adserver container tags using the existing variables which have already been setup. This will allow you to setup basic scripts to conditionally fire the pixels to match the segment. To increase cookie volume, implement separate pixels across all of your CRM emails also.

The tech part is out of the way and now you just need to setup all of the relevant strategies in the DSP including / excluding the newly created CRM segments accordingly.

As new product pages, websites, microsites and CRM email campaigns get created, then the same process needs to take place in order to keep the cookie CRM database updated.

As the cookie database is held within a DSP such as MediaMath, you can deliver the CRM campaigns across ad exchanges, yield optimisers and FBX.


Criteo have certainly become popular over the past few years with marketers. This is mainly due to them being sold into a no hassle site remarketing solution paying on a low risk CPC basis.

What a lot of marketers and CEOs don’t know is that by using a managed site retargeting service like Criteo is not only lazy and inefficient, but it it’s also opening the business to data leakage which could have a damaging effect.

Let’s break those three key points down further:

1. Inefficient way of spending money – Criteo typically charge c 50p CPC yet using a DSP eg. MediaMath and Adacado together, the eCPC rate would be c 5p. Everything works back to an eCPM for display, so on a CTR of 1.5% you’re paying £5 CPM for the managed service via Criteo and only c 70p CPM for the media and dynamic creative service both on a self-service basis. Not only that, 1.5% is quite conservative for a dynamic remarketing CTR so as it goes up, so does Criteo’s eCPM.

2. Lazy – it’s a lot easier to send a monthly invoice to accounts to pay but it does take a bit more intelligence, will power and proactiveness to get a self-service DSP on board and dynamic creative supplier. Display CRM would only be part of someone’s job and they’d typically apply a tight frequency cap and low CPM at the top of the funnel, but as you go lower in the funnel the frequency cap should be increased alongside the CPM as the users become more valuable to your business.

3. Data leakage – allowing a third party to ‘manage’ your data for you is always risky that it could fall in the wrong hands eg. a competitor or data exchange.

Something I haven’t mentioned is scale and volume, although Criteo do have a few exclusive publishers, this doesn’t match to any degree the sheer scale which a DSP such as MediaMath or DoubleClick Bid Manager bring to the table.


Using a DSP on a self service basis lets you set your own rules and strategy structure therefore site remarketing strategies should be completely separate from prospecting. It’s essential to separate them as the KPIs and strategy between prospecting and CRM are so different (similar to brand and generic search) – it’s a classic way of how a lot of ad networks used to over represent the real prospecting display results by including remarketing strategies within prospecting campaign results.

From remarketing site visitors who have bounced, to remarketing high value customers, the set-up should be bespoke to the segment. The further you get in the user journey, the less cookies there will be allowing you to be able to afford to be more aggressive as there’s less risk of budget getting out of control, also the further your customers get in the journey, the more you want to keep them (as their value increases):

Let’s start with remarketing people who have visited but not registered their details – a tight frequency should be implemented and reasonably low CPM.

Dependent on your user journey you should then split out your remarketing strategies by segment (which should all be list in your cookie CRM database) eg. Age, gender, country, product, user level – as the segment becomes more valuable to your business, the frequency cap should be loosened and CPM increased.

As well as the media strategies, the likes of creative and ongoing promotions are crucial for success in this area – it’s not going to help serving your high value customers banners telling them to sign up again or promote an out of date offer. A reminder of what’s currently available, what’s just launched, cross sell, RAF, special offers and what’s coming soon are the basics.

Remember, the harder you work on CRM and generally looking after your customers to avoid them leaving, the more you can afford to pay for new customers which has a positive effect on the incremental volume you can drive (referring to the classic volume price curve).


Claims have been made that you can achieve search performance with display prospecting campaigns via DSPs and some say that display isn’t a DR channel and you shouldn’t just accept post click conversions.

Referring back to a typical volume price curve and the fact that DSPs are now available, it is certainly possible to setup separate DR prospecting display strategies which perform on par with other digital channels such as Facebook and paid search but at the expense of a fraction of the volume.

There are a few ways to achieve this:

1. Setup separate ‘spend’ strategies where you’re throttling the CPM as low as 15 cent.

2. Auto optimising on post click conversions only and the algorithm will only spend if it thinks it will hit the CPA

3. Look at all post click data from previous DSP prospecting and remarketing strategies setting up separate ‘spend’ strategies across areas eg. OS, browser, time of day, publisher….. which index highly / placements which hit your CPA goal.

Just because something is trackable ie. post view conversions, it doesn’t mean that you should accept them all. Also quite a lot of businesses simply don’t have the infrastructure to manage log level data and therefore be able to ignore post view conversions which weren’t even viewed.

Due to various strict policies which both Google and Facebook enforce such as gambling regulation, certain industry sectors would find display to be more valuable as display has minimal restrictions on what you can and can’t do.

There is also a case of ‘why bother with display DR’ knowing that you have to invest in decent DR creative, have access to a DSP and a display adserver and at the end of it all you get minimal conversions compared to other digital channels. My take on it is that it’s a growth area and volume will only increase in the future. Also having a multi channel media strategy is more powerful than focusing on the status quo.

Although many CEOs continue to unnecessarily obsess with banner ads because it’s the most visual way of demonstrating your product creative wise, if as a business you want to stay ahead of the game, you need to seek out all viable opportunities.


Customer lifetime value / ARPU can look dramatically different across the full customer database therefore the ideal scenario would be to optimise only towards high value customers (lookalike targeting).

All you need is a DSP (due to reach, inventory aggregation and optimisation algorithm (which has a brain visualisation report)) as well as having the ability to implement the DSP pixel on a page where your high value customers visit (either conditionally firing a pixel based on variables or simply a premium product page – min. of 200 total global pixel fires a day required recommended)). Once the prospecting strategy sets off, the algorithm will reward the campaign with a conversion event every time someone views / clicks on a banner and then fits within the relevant criteria where the pixel would fire. The algorithm will then deliver prospecting ads that fit within the same internet surfing behaviour and demographic as your high value customers.

There are over 350 million websites in the world and when you duplicate that number to take into account sections of sites, browser, OS, time of day, device, day of week, above / below the fold, internet speed, country, ad format then that’s a hell of a lot of data for even a team of people to manage and manually optimise against for one client throughout the day, which is why an advanced algorithm which has an auto optimisation feature is recommended.

By looking at the brain visualisation report there is the option to also set-up separate manual strategies where you can view a list of targeting options which correlate with your high value customers on an index value basis.

As we know, display is a mixture between BR and DR and falls higher up the funnel than affiliates and generic search so when it comes to looking at ROI and optimising, it would be a good idea to firstly not accept any impressions which have not been ‘in view’ and then accept 1 to 7 days post view and 30 days post click conversion – the whole concept here is to deliver prospecting banners to the right person at the right time and in the right environment.

If you’re interested to find out more, get in touch with MediaMath who offer the best service when it comes to this strategy.