The primary function of a marketing intelligence platform is to provide insights for growth by connecting effort with outcome at granular and aggregate levels.
When marketers achieve this, they achieve a number of incredibly important things.
With a MIP, marketers can slice and dice their data like never before — like calculating CAC by creative asset type. Using a marketing intelligence platform allows brands to connect data that otherwise would never connect: dimensions that might exist just in your internal customer segmentation models. For example, a prospect might convert into a customer via a campaign targeted at luxury buyers, but actually buy a product focused more on utility. Knowing which customers respond to which messages helps brands communicate in smarter ways to customers in order to maximize profitability.
So … what’s required for marketing intelligence?
In broad strokes, there are three major components, each of which is composed of multiple engines. Most of this is invisible to everyday users: they spend their time in the reporting and visualization modules and it just works, delivering the results they need. But without all of these components, a MIP cannot adequately perform its tasks.
Unifying marketing data starts with clean marketing data.
Creative and campaign names can easily proliferate to chaos. High-volume modern marketing organizations work with dozens of marketing technology vendors, multiple marketing agencies and/or partners, and easily upwards of thirty paid marketing channels. A critical step is getting campaign and creative names correct and consistent from the very beginning. Deciding on a common link structure and taxonomy and sticking to it is essential.
Data governance maintains nomenclature sanity, which aids in data unification, normalization, and combining, as well as in certain types of attribution. And it doesn’t just happen at marketing data’s final destination, but rather at its source.
Getting this right enables maximal granularity of data and trackability of marketing impact.
This can’t be accomplished reliably at scale via manual processes. Automating naming and link taxonomy and monitoring of usage ensure this becomes standard operating procedure.
Getting data is one thing. Making it usable and relatable so you can see trends and insights is another entirely.
So a marketing intelligence platform has to run data quality assurance, cleaning and organizing the data. Fraud prevention is a major aspect here: what data do we trust is legitimate human activity, and what data is automated, bot-driven, fraudulent, or otherwise unwanted?
But QA isn’t enough.
Different marketing systems and platforms have different terms and varying methods of measuring prospect engagement and activity, so normalizing and standardizing data is an essential component of unification.
This needs to happen on both the activity/spend/campaign side (the “left hand”) and the attribution/results/conversions (the “right hand”) side. Both input data – effort that marketers are exerting such as campaigns they’re creating or emails they’re sending – and output data – the results that marketers see due to their efforts such as opens, clicks, conversions, and sales – need to be unified and normalized.
Point solutions that simply try to aggregate the left hand or attribute the right hand invariably cannot provide the granular insights marketers need to optimize. Knowing your results is good. Knowing precisely how you achieved them is gold.
Other key components of the data processing step include data enrichment and identity stitching.
More channels and more tools equals more data. As the global datasphere grows to 175 zettabytes by 2025, more and more of that has become addressable by marketers.
Once you have all the cleansed and normalized data, accurate attribution can happen. Now we’re connecting inputs with outputs.
That includes standard last-click or last-touch as well as view-through attribution, plus viewability of assists in a multi-touch attribution model. Re-attribution for lapsed and renewed customers and/or users happens here, as does event attribution: what caused this existing customer to purchase a particular item at a particular time?
In the mobile app world, uninstall tracking and reinstall tracking happens in the attribution engine as well.
Data ingestion is just the first step.
Once you’ve unified and normalized your marketing inputs as well as your marketing outputs, and attributed causation of those results, you need to combine the data intelligently to derive accurate data on conversion, cost of customer acquisition, and ultimately return on investment. That can happen via a unique identifier on a customer, via a deterministic attributed conversion, or via a probabilistic attributed conversion.
But it’s not just about getting simple campaign ROI.
Dimensionality is a key criterion to enabling deeper insights and optimization.
A MIP can ingest and relate data points that are different in granularity and that therefore cannot easily be matched one for one.
One simple example: aggregated marketing campaign data and customer/user-level data. Another: creative assets that are different in size, format, or even technological execution, but that share similar tone, messaging, and belong to the same campaign. Yet another: cost data connected to digital ad campaigns that is not visible in ad network reporting, such as agency fees, registration fees, or processing fees.
With dimensionality, marketers can understand campaign cross-over: prospects who received a luxury offer, for instance, who converted to an economy line. They can also get CAC and ROI per creative assets that are similar in tone and message, not just individual iterations of those assets.
Data needs to be accessible and understandable to be of any value. Reporting and visualization summarize data and tell multiple teams what is happening.
The first customer is the marketing team, of course. But the business intelligence team, the finance team, the C-suite, and others also need to know — at varying levels of aggregation and granularity — what’s going on.
That includes tables, charts, and dashboards, of course, but also customizable reports, pivoting, cohorts, plus extensible reporting frameworks for high-level customers.
Unifying data is a means, not an end. It powers actionable intelligence for profitable growth. In fact, that is precisely CMO’s top priority for their data in 2019 :
Unified and combined “left and right hand” data in a marketing intelligence platform unlocks actual ROI and true CAC across all marketing activities. That’s exactly what unleashes marketers to find pockets of profitable growth: understanding which activities will unlock the highest potential ROI.
Actionable insights include benchmarking both against your own past performance and current broader market conditions, trend analysis, and forecasting for key performance indicators like ROI, LTV, budgets, and CAC.
Higher-level insights come from simulations and testing, and a marketing intelligence platform also delivers recommendations based on both first-party data as well as third-party data. These include lift metrics, media mix modeling recommendations, and fraud analysis.
A system that only ingests data is not useful.
Marketers might need integrations with other measurement systems as well as marketing action platforms like email and push notifications. Marketing intelligence platforms also send data to internal BI systems via either API or S3 data dumps for further processing, analysis, or integration.
And, of course, partner mediation in terms of postbacks on installs, events, or purchases is critical.
A marketing intelligence platform cannot unify marketing data without speaking to dozens of different kinds of systems, and thousands of individual platforms. And that communication is often bi-directional, with enrichment happening from multiple sources at multiple stages.
Even though modern scientific marketing is not a set-it-and-forget-it activity, marketers increasingly need to be able to automate actions within set parameters.
That begins with customizable alerts for when campaigns fail to hit or exceed parameters.
It includes automated creation and distribution of audiences for retargeting, look-alike campaigns, or suppression lists. It also includes built-in on-by-default configurable mitigation of fraud, along with both whitelisting and blacklisting of sources and publishers in paid media campaigns.
Marketing intelligence must find and minimize fraud, since it simultaneously wastes spending, decreases ROI, increases CAC, and skews analytics. Fraud management starts with detection but includes automated abatement. Plus, it gives marketers the ability to drill down to the sources of fraud in order to take further action.
Finally, at higher levels of functionality, a MIP automates bids and buys for ad campaigns at scale, enabling marketers to make high-level allocation decisions that are automatically optimized via intelligent, learning systems. Successful campaigns continue and grow; campaigns where CAC increases unsustainably shrink or get shut down entirely.