By Tom Grubb, Chief Strategy Officer, Digital Pi
What can you say about marketing analytics that hasn’t already been said? Plenty when you know how many well-intentioned marketers too often misrepresent, misread, misunderstand, miscalculate and flat-out misuse – or don’t use marketing analytics at all. Considering the amount of time and money companies invest in people, technology and marketing programs one would think that getting marketing analytics right and putting them to work to improve ROI would be a top priority for every organization right up to the CEO. Even the companies that make marketing analytics a top priority often fail to set the right objectives, ask the right questions, and drive adoption across business functions. The journey to transforming your organization to embrace marketing analytics as the common language of marketing isn’t a journey at all: it’s a mission that requires a leader who’s not afraid to take on this formidable challenge.
In my work consulting with Marketo customers I’ve seen how marketing analytics champions can become a lightning rod for criticism from all sides for many reasons. The area of analytics that causes the most problems for those championing analytics is revenue attribution reporting – the method of tying marketing performance to revenue in order to use pipeline and revenue as a guide for optimizing marketing. In plain speak, the question can be framed as “show me which programs or tactics had the most/least influence on pipeline created or revenue won.” The good news for Marketo customers is this: the Marketo platform is designed to collect and manage data in a way that lends itself to revenue attribution reporting. Before I dive into Marketo’s revenue attribution model, we need to consider a very different approach to revenue attribution that I encounter in various forms: single source revenue attribution.
For this article, I am focusing on B2B companies. Single source attribution says, we can attribute the entire credit for an opportunity to one buyer – say, Sarah Smith, the Director of Purchasing. If lead source is the one and only attribution we apply, and we say Sarah is the person in the deal who counts the most, then we look at Sarah’s lead source – trade show, and apportion all of the credit for revenue to the trade show that acquired Sarah’s name in our database – great job marketing! Or alternatively, assume instead a sales person found Sarah on LinkedIn and added her name to the database. In that case, all of the credit for the deal would go to Sales for finding Sarah. There are lots of ways to declare single-source credit for a deal, from acquisition program to last response, to home-grown algorithms. The overarching goal for single source attribution is usually to determine whether sales or marketing sourced a deal, so management can allocate resources based on which function is sourcing the most pipeline or revenue. Dividing the world into sales or marketing revenue attribution has great appeal for obvious reasons, but it can lead to a distorted picture of how marketing and sales influence pipeline and revenue creation.
The example above is built on the idea that a business can declare one person as the entire reason a deal happened, to the exclusion of all other people and their engagement that influenced a deal. In reality, most B2B deals involve more than one contact from the buyer, often defined by their role in the decision such as influencer, or primary. Depending on the product or service being sold, a seller might engage with a few people or many dozens. Let’s say for our purposes we’re focused on deals that average five to ten buyer contacts for any opportunity. In order to get a complete accounting for the role marketing played in any opportunity, we have to consider all of the people associated with the opportunity and understand all of the marketing programs they engaged in – like attending webinars, visiting a booth at a trade show – any and all buyer responses to a marketing stimulus.
In simple terms, B2B marketers are responsible for 1) adding new names to the marketing database and 2) engaging them through marketing programs. If marketing is successful in pursuing those core missions, some of those people should make their way onto opportunities, some of whom should eventually convert to revenue. The big picture looks like this: marketing adds new names acquired by a number of programs/tactics, and marketing engages people through a range of tactics/programs that change frequently. Compared with single-sourced attribution, the world of multi-touch attribution can get very complicated very quickly.
As I stated earlier, the Marketo platform is designed to collect and manage data in a way that lends itself to revenue attribution reporting. Understanding this architecture is the proverbial key to the universe when it comes to getting the full value from your Marketo investment and getting to good analytics. If you already know Marketo, read on anyway because this simplified explanation can come in handy when you have to explain Marketo to the un-indoctrinated.
Marketo channels are your marketing tactics defined as templates – Webinar, Trade show, Online Advertising and so on. For each channel you must define the actions someone can take – these are called member statuses in Marketo. As an example, the Webinar channel typically includes: Invited, Registered, No Show, Attended, and Attended On-demand. You must also state which member statuses you consider a marketing success. Using Webinar as an example again, you would definitely say people who attend your webinar reached success, but when/if someone only registers – is that a success? It depends if you consider someone registering for a webinar a marketing win; some do, some don’t. Marketo declares all successes are created equal – that is, a webinar success has equal value to a white paper download. This prevents marketers from trying to be too clever at assuming we know which engagement influenced someone more than another. There are lots of attribution models out there, seemingly more every day – this article is devoted to Marketo’s attribution model.
There’s one last component you need to understand to get the Marketo architecture religion: Marketo programs. you must create a program for every webinar, trade show, ad campaign, etc. that you plan to run. Those Marketo channels I explained above? Here’s where they come into play. Every time you create a program, Marketo asks you which Marketo channel to use as the template for the program. If you’re creating a program for a webinar, you pick the Webinar channel; for a trade show – the Tradeshow channel and so-on. After you create a program, you can start adding names to the program as members. Those member statuses, like Registered and Attended? For every member in a program, you can (and must) build the logic in Marketo to change their member status when they do whatever it is that maps to a member status. For example, someone in a webinar would start with member status is “Invited when you add them to the program, then when you send them an invite the member status needs to change to “invited,” when they register it changes to “Registered” and so on.
How you define and use your Marketo channels, member statuses, and which member statuses are successes is the key to defining how you will measure, analyze and optimize your marketing based on marketing analytics. Salesforce has similar architecture design concepts that map to Marketo: Marketo Channels = SFDC Campaign Type, Marketo Channel Member Statuses = SFDC Campaign Member Statuses, and Marketo Success = SFDC Responded. The big differences are, SFDC campaign types are not templatized for re-use, and SFDC does not support the idea of success defined according to each unique channel member status. These seemingly small differences are what give Marketo a big edge in collecting and storing data for reporting. Now you have the essential building blocks to get your Marketo analytics right. I cannot overstate how important it is to get these right, nor how often we see Marketo customers miss on these important elements.
Think of marketing as a betting game. You start with a plan, placing your bets on the calendar across marketing tactics and campaigns. You bet some of your budget on webinars, some on newsletters, trade shows, paid online advertising – spreading your bets across multiple marketing channels and dates. The dates approach, the programs run, things happen (or don’t). Some programs bring in new names, some engage names already Marketo. How do you know which bets paid off, and which ones didn’t?
Let’s stick with the assumption that on average, an opportunity has five to ten contacts we engage for any opportunity. Instead of saying any one person or one action gets all the credit for an opportunity, we spread the credit across all of the people associated to an opportunity. If there are five contacts on the deal, we must examine the marketing history of all five contacts to determine the impact marketing had on the deal. If the marketing history for those five contacts shows they were members in twenty Marketo programs, Marketo would examine all of their marketing history in all twenty programs to determine how much credit each program receives for influencing the deal.
Here’s the big idea: it takes multiple people acquired and engaged by multiple programs across multiple tactics, played out over time to make a deal happen. You have to connect opportunities to people and drill into their marketing history to connect marketing with pipeline and revenue. Many struggle to fully grasp this idea, so I will put this another way: marketing doesn’t invite opportunities to webinars, or talk with opportunities at the trade show booth. If the people they invite to webinars or talk with at the trade show booth eventually become associated to opportunities, Marketo can find the connection between marketing and the opportunity through their actions by First Touch, and Multi-Touch attribution.
If the goal of the program was to acquire new names, the measure of success will be how many new names the program added to Marketo. For example, if you setup a program called “March 2018 Google PPC” to collect new names brought in by your Google ad campaigns, and the program added 100 new names in a month, the “March 2018 Google PPC” was a good bet for Marketing to the extent that it added 100 new names. If you setup your program correctly, every name added to Marketo that originated from “March 2018 Google PPC” program will have the program name “March 2018 Google PPC” contained in the Acquisition Program field on the lead record. This field is a Marketo default field established when you setup Marketo, and an important one at that. If you want to know which marketing tactics/programs performed best to add names to the database, this is the field Marketo uses to answer the question.
In the Marketo Revenue attribution model, Marketo calls this First Touch attribution – the program that gets credit for adding a person to the database. In hindsight they probably should have named it “Program Source” or something like that so people wouldn’t misinterpret it to mean the first program that engaged a person. Marketo acquisition program is similar to a lead source because it addresses the question: where did this person come from? But lead source and acquisition program are different – though there is often uniformity between the two, as in the case when a trade show program acquires a name, we can infer the lead source is trade show from the fact that the program that acquired the name is built on the Marketo trade show Channel. In Marketo, you need to have lead sources and acquisition program values assigned every time, and consistently – even when a name is added by a sales person from the CRM. That’s right, you should create a Marketo channel, call it Sales Generated, and create a Marketo program on the channel to collect new names and assign the First Touch credit for new names added by way of the CRM. That way, you can compare the ROI on acquiring names from the CRM with marketing tactics like Online Advertising or Tradeshows.
If the goal of the program was to engage names already in the database to prompt them to take a meaningful action – like attend a webinar – every person reaching success in a program puts that program on the boards for consideration toward sharing in the credit for a deal. Think of Marketo as a board game again. You play the game by placing your marketing bets on the calendar with programs and tactics. You score early wins when people reach success in your programs (attend, download, etc.). Those successes translate to pipeline and revenue influence. If the object of the game is revenue, the winning strategy is to drive true engagement all the time, where the engagement (Marketo program success) translates to pipeline and revenue influence. Note I said influenced, and not created. Marketing does not create qualified opportunities – marketing engages people some of whom will make it to opportunities. Here is a simple example of a multi-touch scenario.
Sue the prospect attends a webinar and downloads a white paper. Bill the prospect downloads the same white paper and attends a live event. Suppose Sue and Bill get added to a $10K opportunity, Marketo will look at Sue and Bill’s marketing history and tally up their successes:
That makes four successes total across three programs. Sue and Bill were both in the White Paper program so that program gets 50% of the credit, the other two programs each take 25%.
If the $10K opportunity makes it to closed won, the revenue won amount will be applied by the percentage portion that each program received: $5K to White Paper program, $2.5K each to Live Event and Webinar programs.
Marketo also splits First-Touch across people, so in this example Marketo would determine the acquisition program for each person and split the first-touch credit accordingly. This was a simple example, as you can imagine the more people involved in a deal, and the more time / engagement that goes by, the more complicated revenue attribution gets. Depending which reporting tool you use, Marketo gives you options to filter and calculate data using different assumptions, including whether to count program successes up to opportunity created date, or through opportunity closed date.
Now that you understand the basics of Marketo revenue attribution, you must learn where and how to report revenue attribution. Before I dive into that, you need to understand how Marketo calculates the numbers and where those calculations are stored. Those calculations are performed nightly and stored in an application and database separate from your Marketo application and data. Here are the analytics tools Marketo sells to access and analyze the performance data:
Marketo Advanced Reports Builder (ARB) – formerly named Revenue Cycle Explorer (RCE). ARB is a BI tool that uses an Excel-like pivot table metaphor that supports drag and drop to assemble reports, show them as charts and dashboards and a lot more. When you launch ARB from Marketo, you’re firing up a separate program. Don’t go looking for ARB in your Marketo UI, it’s still referenced by its old name Revenue Explorer.
Program Analyzer: An x-y axis chart that lets you pick and choose plots using revenue attribution field and values. Interestingly, this tool has a few calculations that you don’t see anywhere else, like cost per MQL.
Marketo Performance Insights (MPI): this is a new tool offered by Marketo that is purpose-built to make it easy to see program success and revenue attribution packaged in the Marketo Insights UI.
Where ARB gives you the kitchen sink to build anything you want the way you want, MPI provides the essential program performance reports pre-built, with some flexibility provided in the form of drop down filters, and export features. What you trade-off for the complete flexibility you get in ARB, you gain in rapid time to productivity and from my limited experienced so far with MPI, some real-time performance gains on refreshing charts and tables.
Marketo was early to the marketing attribution game to realize the impact contact attach rate with opportunities (or lack thereof) has on revenue attribution reports. If sales doesn’t add anyone to an opportunity, there’s nobody available for Marketo to consider for marketing influence. So Marketo gives you a configurable setting that determines how it will identify people to associate with opportunities:
Explicit: only count the contact roles associated with opportunities (this is SFDC’s one and only way to make the people connection for its pipeline influence report)
Hybrid: look for a contact role on the opportunity; if it doesn’t find at least one, look at the opportunity company, find the account contacts with and without roles, and consider all of them for influence
Implicit: Look at the opportunity company, find the account contacts with and without roles, consider all of them for influence. Think of this as account-based influence.
When you change the setting, it takes effect in the next overnight data calculation. You can run the numbers under all three settings over the course of three days. Given the time lag, it’s a good idea to plan your reporting well in advance as much as possible.
If you design your Marketo correctly and apply rigor to process and data every day, you will get revenue analytics that tell you a lot about your marketing that will help you be a smart marketer. If you don’t, you may expose yourself to revenue analytics that will tell an incomplete, or inaccurate story – and you may not even realize it. Even one bad list import can have a detrimental impact on your reports. Here is my attempt to list the data and process areas you need to consider:
There may well be more I have not listed here. What’s the biggest room in the world? Room for error when it comes to revenue analytics (not just Marketo, any revenue analytics). It is possible to control these challenges, but you need to be aware of them, monitor for them, and get everyone on board to keep your data clean.
It isn’t enough to understand how this all works. It isn’t enough to get your data right and keep it right. If you understand Marketo revenue attribution, and you get your data right, and you build the standard set of reports -- you still have the most formidable challenge ahead of you: getting your organization to understand, buy into, and use the reports to continuously optimize marketing ROI. If marketing needs its own language – and it does – it’s marketing analytics. There’s a lot more to making marketing analytics the common language of marketing in an organization than I can cover here. It’s a subject I have a lot to say about in the forthcoming May 2018 issue of Applied Marketing Analytics: "How organizations can establish marketing analytics as the common business language to drive continuous improvement."
Originally posted on LinkedIn on April 24, 2018
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.