Measurement of performance marketing necessarily has to be custom to fit the needs of each individual business, but the following sections outline some key considerations to put you on the path toward properly measuring performance marketing for your business.
Contents
Product and Marketing Alignment
For both simplification and to ensure consistency across the organization, product and performance marketing teams should create as much alignment in measurement between teams as possible. This means using the same infrastructure if possible, agreeing on data definitions and nomenclature. These teams have different objectives, so it will often not make sense for them to use precisely the same metrics to track progress, but it’s imperative that the teams do not have different measurements with the same name as this will inevitably cause confusion across the organization.
Cross Platform Tracking
Tracking just on web or just on mobile keeps things pretty simple, but most businesses need to be able to do both and accurately.
There are many viable systems for attribution on web, but on mobile you are limited to companies that work properly with SANs (self attributing networks) like Facebook, Google and Snap. Even if web is the majority of your business, you will need to either use two systems to track, or rely on a tracking solution that is mobile first with additional features for web.
User Based Tracking
Attribution by its nature is device based, which is functional but can result in long term deficiencies in accurate measurement over the long term. Moving to user based measurement will prevent miscounting due to
- A user interacting with multiple devices at once
- A user changing devices during their lifecycle
- Multiple users on a single device
Spend and Attribution Matching
Special attention should be given to how the cost of advertising is being mapped to attribution data to accurately understand metrics like CAC and ROAS. It is key to know if your cost information is accurate at the campaign level, the country level, or the creative level. How granular you are able to pull cost data from any particular channel will impact how you structure your marketing campaigns, and how you build out dashboards.
KPI and the Funnel
Before building any reports, the organization needs to align on the KPI that most closely aligns performance marketing with the business goal. Mapping out the funnel of steps a user follows to get to that KPI will result in a list of all of the constituent metrics of that KPI that will help to describe opportunities in optimizing the KPI.
Projection of LTV and ROAS
How complex your models need to be to accurately project LTV will depend on the potential range of user values. Products with a single subscription will need very simple models, while products with potentially infinite spend from a single user will need skilled data scientists and complex machine learning models to make the right budgeting decision. A simple test is to run a correlation analysis between early and late LTV by marketing campaign cohort. If the correlation is high, you can get away with something simple like a consistent factor applied to all cohorts.
Measurement Levels
Once you have campaigns running there will be a lot of data coming through and I think it’s helpful to have a framework for thinking about this granularity.
| Level | Uses | Examples |
|---|---|---|
| Low Level | This is where the media buyer lives and optimizes | Campaign, Country, Placement, Creative |
| Mid Level | This is where strategy is set and is most useful for stakeholders who need some level of detail | Network, Platform, Region, Optimization Type |
| High Level | Executive level reporting | By Date |
Dashboarding
Dashboards will need to be built for the performance marketing KPI and the constituent metrics for the high, mid, and low level cuts of data. For the high and mid level, charts are the most appropriate form. For the low level charts are useful and a table view will provide the best usability for the UA manager.
Marginal Impact
Managing performance marketing based on the average performance of each campaign is both useful and relatively simple to understand, but it’s important to remember that this does not necessarily result in the optimal mix. The marginal performance of each campaign will be different and to optimize mix, it is the marginal performance that is important. Measuring this is not easy, but there is value in understanding marginal returns per campaign in order to arrive at the most efficient mix.
Uplift and Incrementality
Device and user attribution has been at the core of performance marketing for decades now, but it is important to remember that this is not the end all be all of measurement. It is the only real way to measure things like LTV and ROAS, but can have real downsides. Due to the high level of investment required to do this properly, uplift and incrementality should be approached when
- You are spending a meaningful amount of budget on search for brand keywords
- You have budget going to channels that are very leaky (e.g. influencers, marketing with QR codes)
- You are doing any reengagement marketing (NEVER use attribution to measure reengagement campaigns)
- Most customers convert after multiple touchpoints
Final Thoughts
You will not be able to do everything at once. What is important in the beginning is to get the basics in place, make sure the vendors you choose will be able to get you to where you want to go, and that you have a roadmap to continue to get your measurement to most accurately reflect your impact on company profit and to provide as much transparency as possible.
Thanks for the article!