By Rob Webster, Chief Strategy Officer, Crimtan.
There have been two broad themes in programmatic over the last year or two. Firstly, that running a programmatic campaign is fairly easy, so paying a supplier to do it does not require a high margin (hidden or otherwise). The second theme is around the number of detailed issues that programmatic needs to resolve – a short headline list of which would include: performance, viewability, fraud, brand safety, attribution, device/cross device, programmatic creative, data usage, planning, machine learning, GDPR and many more.
So how can we reconcile these statements? Well it’s certainly true that it doesn’t take much to run a programmatic campaign, either in house or with a programmatic supplier. Graduates can be trained in weeks on how to use a DSP and set to work. However, anyone doing simple buying on an exchange, even if they can get their performance metrics right, would almost certainly suffer from awful viewability, brand safety and fraud concerns before even looking at more advanced topics such as attribution, machine learning and data usage.
Some problems can be solved through resource. By adding more manpower to a campaign you can apply more techniques and optimise more precisely. More resource means more testing of new areas, more optimisation and more analysis – all of which leads to better results and improved campaign performance. The trick is to apply the right amount of resource to the campaign. Broadly this can be done as a function of spend though some campaigns types need, and deserve, more resource than others.
Even though resource is often thrown at programmatic issues, and programmatic exchange buying is over 10 years old, only a very small proportion of advertisers’ programmatic systems get close to addressing all of the concerns to an acceptable level. In fact, defining “acceptable level” is a crucial point to be considered. Almost any programmatic campaign can use an almost unlimited amount of manpower, so it’s important to add expertise to judge what to improve and how to improve it.
Imagine an advertiser with spend on the low side. This requires a very limited amount of resource to run, let’s say half the time of an inexperienced programmatic buyer’s time. However it will also need to be looked over from time to time by an expert and given a health check on areas such as viewability, brand safety, fraud, strategy and so on. The expert might advise a change in KPI’s, a shift in bids, some tactical changes to the strategy or optimisation and many other pieces of advice. The expert would also advise on the technology needed. A much larger campaign would need more resource but also a much larger amount of expertise and more advanced technology (perhaps specialist technology for DMP, attribution, fraud and creative).
So, as an advertiser, you understand that you need resource, expertise and technology – but you also need creative, data and media. And how much should you spend on each as a percentage? Recent articles have come up with a figure of around 60% going to media, with broadly 10- 15% on service, 10-15% on Technology and 10-15% on Data. The exact amount being dependent on a number of variables beyond the scope of this article (including industry, strategy and scale). I have left creative out of these costs because it can be part of a wider conversation and a different budget, though I would recommend spending up to as much as 20% of media spend on creative.
Lastly then, this is where we have to be nimble. The world of programmatic is constantly changing: new opportunities emerge (e.g. video, native, dynamic creative, DOOH, OTT TV), new technology solutions to existing issues like attribution are built and, inevitably, new challenges appear. So expertise working alongside a nimble and agile culture are vital. Expertise and resource can be wasted if it is married to rigid structures built for the challenges of the past. As programmatic activity scales, teamwork and shared learning become ever more important. This area of nimbleness is worth an article all on its own and I can’t underestimate its importance.
So bringing things back to the start of this article, what are the implications? Yes running programmatic can be easy and it should be transparent. However, to be done properly it needs the right amount of resource, expertise and technology in a nimble environment. This dynamic helps us answer some of today’s questions:
Q Should I run programmatic in house?
A If you can provide the right amount of resource, expertise (which could be outsourced) and technology (which can be bought in) in the right environment, then, yes. Many advertisers though will need to look externally for expertise and possibly a programmatic culture.
Q How will machine learning impact programmatic. Will all jobs be automated?
A Much of the “resource” part of this article can, and will, be automated in future, but not all. Equally, more and more of the tactical optimisation will be done by machines. However we are decades away from automating the expertise and the programmatic culture required to run best in breed operations. Machines won’t fully take over until the arrival of true artificial intelligence that can mimic a programmatic expert’s strategic knowledge and understanding of the space. This means that entry level jobs will become scarcer and we need advanced training programs to get young professionals more quickly to a position where they can add strategic expertise.
This then leaves the long list of problems we started with: performance, viewability, fraud, brand safety, attribution, device/cross device, programmatic creative, data usage, planning, machine learning, GDPR and many more. With the right structure in place all these problems and the new ones that will inevitably come up are very solvable. We are currently in a good place where programmatic is easily accessible but is complicated enough and has enough value to reward excellent work. Such a dynamic is hardly new – for decades marketing has been easy to grasp, tough to master but achievable with the right expertise.
The same dynamic is true of programmatic, and advertisers will succeed by backing expertise and a nimble, agile approach. As always then, the more things change the more they stay the same. The only real change that programmatic has brought about is speed – where speed of optimisation, speed of change and speed of opportunity are exponentially faster than in Don Drapers day – so, you really do need to be nimble!