#1 on Inc. Magazine’s List of 50 Best Websites for Entrepreneurs
#2 on Forbes List of 100 Best Websites for Entrepreneurs

Marketing Attribution: Creating a Growth Engine at Salesforce, Zendesk and Slack




I recently sat down with Bill Macaitis. I’ve been a big fan of Bill’s since we brought him in as CMO at Zendesk. He has since gone on to be CMO at Slack. In an age where product gets most of the attention, Bill is the kind of marketer founders crave. He knows exactly what impact his marketing efforts are having. He makes it easy for a founder to expand his budget. And he’s able to build a predictable growth engine.

How does he do this? The key, Bill would say, is marketing attribution–being able to evaluate how each marketing channel is contributing to growth. Many startups haven’t considered just how difficult it is to accurately track ad and marketing performance (the last ad clicked is not the whole story.) Or, many startups believe the analytics that come with marketing attribution is better left for when a company reaches scale. Bill firmly believes it needs to be put in place early. Why start any marketing if you can’t measure what’s working? In this post, we share ten years worth of learning and experiences Bill has amassed with marketing attribution–what it is, why startups should do it, and how they can best put it into place early in their company’s life.

What is Marketing Attribution

In social psychology, attribution is the process of inferring the causes of events or behaviors. Marketing attribution has been defined as “the science of assigning credit or allocating dollars from a sale to the marketing touchpoints that a customer was exposed to prior to their purchase.” Bill Macaitis, CMO of Slack and formerly CMO at Zendesk has a much simpler view:

Marketing Attribution means: are you spending your money wisely?

A simplified example: let’s say you have a customer who first saw your display ad but didn’t click. Weeks later they saw your Facebook ad, clicked through and read several of your blog posts. Still weeks later, they clicked on one of your Google ad words and ended up signing up for a free trial. How do you value each of these marketing channels? If the display ad hadn’t been seen, would it have mattered? Would they have found their way to your site without the Google ad word? Which blog posts resulted in the highest value signups?

A simplistic last-touch model would give all the value to the ad word and undervalue earlier touchpoints. A first-click model would give too much credit to the first click, the Facebook ad. Equally or linearly spreading the value across all channels would just be guessing. To get a much more accurate view of each channel’s role and be able to optimize your spend, you need an attribution model that analyzes all of your data–who viewed what, who clicked what and what actions did they take–and algorithmically determines the impact of each touchpoint.

Why You Need Marketing Attribution

As you ramp up marketing and start using different channels, you need a more sophisticated approach than simple click tracking to follow your buyer’s journey and know what’s working, what’s needed when, and where your spend is paying off. This is especially important for a more considered purchase, such as a B2B sale where it can take months and five, ten or more touches across various channels before your customer takes their first action with your company.

Bill has been building marketing attribution systems for almost 10 years–first at Salesforce, then at Zendesk, and now at Slack. At each company, marketing has become a huge competitive advantage. These companies, once they found product-market fit, were able to generate hyper-growth in large part because they knew they were spending their money efficiently and could thus scale up aggressively. It’s meant the difference between 40% to 50% growth per year, to over 400% growth per year.

But, how does a startup approach this? Don’t you need complex systems and data scientists to get this right? Don’t you need a large marketing spend? Not necessarily. Luckily, we live in an age when marketers, even at startups, have more data at their disposal than at any other time.

A New Era of Marketing Attribution

It used to be all we could really do to understand the impact of our marketing spend was to track the first or last click to a paid conversion event using simple referral links. This worked well, for a while, especially when SEM was the majority of most companies’ digital spend. But, as digital channels expanded we needed a way to recognize the influence of these other channels. Along came rules-based attribution which allowed multiple channels to be tracked and weighted. But, the accuracy of a rules-based approach was limited because the values of each channel were inputted by the marketer. Intuition was still driving what each channel was worth, rather than data.

Today, algorithmic attribution has become the best practice for data-driven marketers and companies. We can now utilize all the available data collection, tools and models to take in all different touch points and make predictive, algorithmic attributions. When set up properly, we can track each touch point and all downstream funnel metrics. And by weighting proportionally across a very large data set, we can determine with much more accuracy and precision what should get the credit–including both online, offline, performance-based and brand advertising.

It’s not perfect, and it’s not easy. It gets difficult with word of mouth referrals, dark social, and other “hidden touches.” But, it drives a much deeper understanding of the buyer’s journey and which of your marketing efforts are paying off.

People make passionate arguments about first or last click attribution. When it’s my budget and company on the line, I go with the data and science every time. — Bill Macaitis, Slack CMO

When is it too Early for Marketing Attribution?

It doesn’t make sense to scale any marketing before you have product-market fit. But, once you have this and you’re starting to think about your marketing mix, you need marketing attribution. You don’t want to be in a position where you’re starting to spend money and you don’t have a way to track it. Knowing what’s working is too important to the growth of your business to push it off until you have a larger team and larger marketing budget, for it will guide your overall investment decisions, your staffing, and how you resource the team.

If you’re using qualitative and subjective assessments instead of data to make decisions you can end up in a really frustrating situation. Without data, you are forced into ivory tower type debates over buying this billboard, or that display ad, or which homepage to go with, and you’ll be guessing based on your own personal preferences. Even when you are early in your company’s growth, you are best served letting users decide what resonates most by tracking their touch points and outcomes.

Of course, putting in a system for algorithmic attribution requires a significant investment of time, money, and people. And that can be tough for a startup where resources are tight. But, there is a way to start smaller. And, relative to the results you will achieve from being able to optimize your growth levers from the beginning, you will see an incredible ROI on this investment.

A Guide For Implementing Marketing Attribution at a Startup

Once you have a few channels you’re using for marketing (ie. webinars + content marketing + fb ads), you need to get some basic tracking in place. This is the non-sexy groundwork that will end up guiding all of your decisions down the road.

1) Educate Your Team

Because you will need an investment of people and dollars to build your attribution system, and because the data you collect will drive fundamental decisions about how you market to and engage customers, you need to start by educating your internal team. Often in an early tech company there are few, if any, marketing people, and people haven’t wrapped their heads around attribution–what it is, the need for it, how to “fairly” assign credit to different channels, how to account for the longer nurturing cycle required, how to factor in word of mouth, etc. At the same time, companies are rightfully focused on their capital and burn and not looking to make big infrastructure investments in marketing. It comes down to trust–which you need to build very early on and continue building as you share the data.

This is a classic change management process that starts with a lot of education before anyone will be comfortable investing.

People may have very impassioned arguments for why one type of attribution is better than another, but you need for people to understand that marketing doesn’t have to be a guessing or intuition game. Put it into terms they will understand–we are going to collect a large set of data and let the data science determine what channels and campaigns are having what impact.

2) Hire the Right Team

As I shared in my introduction, Bill typically recommends the first person a head of marketing should hire is a marketing operations person. These people are worth their weight in gold. When you’re first starting, this may be more of a generalist role, combining operations, analysis and possibly other functions. Look for people who have background in building out a marketing tech stack with a focus on attribution. Even if they haven’t done attribution specifically, a lot of people who have done SEM or media buying will inherently grasp looking at ROI vs spend and will have some experience with the different viewpoints on attribution. The key is finding people who believe in data driven decisions. Interview questions Bill uses to get at this include:

  1. If we ran three different ads, what would you look at to determine which ad was most successful?
  2. If we spent $1M on advertising this year, how would we know it was successful?  

This person will also need to be a strong partner with your development team to build out the tracking infrastructure and integrations to your data warehouse. They should have a familiarity with databases, and if you’re short on database resources, you may even want someone who can set up some of the early database queries themselves.

Hire this operations, data-focused person first so that you can begin tracking from the very beginning. From there you can start to lay the groundwork for other key teams including product marketing, content & editorial, PR, advertising, lifecycle nurturing, events and customer references. Marketing tends to grow in many directions and can become quickly unwieldy, untrackable, and expensive when you aren’t measuring the impact on growth. You will build trust for future marketing efforts by being able to assess the impact you’re having from the very beginning.

3) Buy or Build

The marketing attribution space is crowded, with plenty of pure play vendors like Convertro (bought by AOL), VisualIQ, Google Analytics (bought Adometry), as well as many smaller players to fit different budgets. You may get along with Google Analytics to start, or some marketing automation tools also come with attribution modules.

At Salesforce, Bill built his own system when they had a lot of custom tracking needs for events they were attending, and there were many fewer vendors to choose from. Now, it’s worth a thorough search of vendors to see if there is a tool to fit your needs. Just the process of evaluating vendors will teach you a tremendous amount about marketing attribution. You’ll want to pay attention to capabilities such as the basis for their algorithm, their method for tracking user level data, whether they have cross device tracking and the strength of their match rate, as well as their pricing and how easy it will be to integrate and operate their tool.

4) Integrate for Deep Funnel Metrics

Marketing needs to form a really strong partnership with your development team and those who manage your data warehouse. For attribution to work and to get the most actionable data, you’ll need to integrate into your data warehouse to track specific customer actions. Don’t stop at a customer signing up for a trial, but follow it through to understand if they converted to a lead, an opportunity, to a signed deal, and what kind of expansion revenue they contributed. Your focus needs to be on your long term, core business metrics. Tying marketing value to short term metrics can lead to the wrong incentives and actions.

5) Start Small

You don’t need to integrate a pure-play, end-to-end tracking system to track ten different channels on day one. If you’re starting early, you probably don’t even have ten channels to track. Start with Google Analytics or another tool you’re already using that may have some tagging and tracking abilities. Track a couple channels, e.g., ad words, some display ads, and maybe choose a couple performance networks. Focus on three metrics you care about the most, such as filling out a registration form, receiving a demo, or becoming a qualified opportunity or signed contract. If you have different sizes or lengths of contacts, pretty quickly you may find a lot of value in tracking the resulting ACV to see if different channels are impacting this key metric. Over time you’ll add more channels and start to track at a more granular level, but for now, keep it simple.

Example #1: Zendesk – Beyond Paid Ads

When Bill was CMO at Zendesk, the market for attribution had evolved such that there were a number of pure play vendors with great offerings. He ran an exhaustive search and they ended up selecting one of the larger vendors. They did a lot of work to integrate into their data warehouse to track across paid display ads, email, paid social, organic social, SEO and SEM. By tracking these channels and individual customer touches they identified a key insight–prospects had to interact with Zendesk ten, twenty, sometimes thirty times before they even became a lead. Knowing this, and measuring the influence display was having during this nurture cycle, they continued to invest in display. Without this level of tracking, they would have underestimated display and potentially cut off or reduced this important channel.

Marketing attribution isn’t just for paid channels. Content marketing can also be tracked to determine what pieces are adding value and when in the customer journey. Zendesk invested heavily in content marketing to reach prospects during their lengthy nurture cycle. They set up tracking to see not just every ad a prospect viewed, but also every piece of content. They were able to track which pieces made a person more likely to buy, which increased the velocity of sales, and which led to bigger deal sizes. With this information they could optimize their content flow and confidently increase spend, knowing every additional dollar they invested in content marketing added to the company’s bottom line.

The results? Zendesk was able to dramatically accelerate lead and revenue growth by understanding the impact on ROI across all of their channels and increasing spend on advertising and content in a smart, strategic way.

Example #2: Slack – Taking it to the Next Level

The way you set up your attribution tracking needs to align with your core business objectives. At Slack, the metric of success is based on the total experience customers have with the brand and offering, and ultimately whether they recommend the service. It’s not just whether they sign up. Add to this that Slack limits the number of registration and signup forms, favoring a delightful experience over contact collection and tracking. Finally, add in a significant focus on word of mouth marketing and offline marketing, including TV ads and podcasts. All of this adds up to a very complex attribution marketing system.

While most startups will not need to start with the multi-layered system Slack has developed, there are tactics they are using that others will find helpful:

  1. Because Slack focuses on a team concept, not individual use, they have developed tracking for deeper team-based funnel metrics including size of team, average ARR per seat and activity across teams.
  2. To track “non clickable” performance ads and brand focused touch points from TV, billboards, magazines, radio and podcast advertising, Bill put in place a robust system with multiple tracking signals including:
    1. Referral codes and coupon codes (e.g., get $50 in credits when you go to’name of podcast’).
    2. Local test markets (advertise in a collection of cities) and compare those cities’ growth rates (visits, leads, aided recall, expansion ARR) against your control group (similar cities or rest of US).
    3. Multiple brand metrics are tracked including aided brand recall, unaided brand recall, visits to website, sentiment and share of voice in addition to traditional funnel metrics (leads/opps/ARR).
    4. Signup survey which lets users choose multiple channels where they had heard about Slack (including offline channels and word of mouth).
    5. Bill is also a big believer in NPS and tracks NPS at both a relational level (would you recommend Slack in general) and transactional level (would you recommend Slack after a specific encounter with a feature or service).
  3. All of this data gets plugged into their attribution model to find which channels are having the most influence on a customer’s ultimate satisfaction with Slack.

Attribution Is More Than Acquisition

Slack is also a great example of how attribution can be used beyond acquisition to value the touchpoints a customer has throughout their lifecycle and how those touch points increase usage of and satisfaction with their product. Because so much of the value of SaaS comes after the sale–the so called “land and expand” approach where you start with a small instance and then work to grow it and expose it to more people–companies everywhere are investing in customer success. Most companies focus this lifecycle marketing on “free channels” like email and in-product messaging. But, paid channels can also play a role with existing customers. And, when you have an attribution system tracking your existing customer marketing and can determine the value of different efforts and spend, you will be able to optimize both paid and “free” channels to have the biggest impact on customer growth and retention.


Attribution is simple in concept, but complex in execution. It’s also a quickly evolving field. Fortunately, there are many great resources to learn from–blog sites, podcasts, vendor sites and conferences, to name a few. We hope this article has started or continued your learning process. Marketing is indeed fuel for the fire. But, don’t spend blindly. Set up the tracking you need to scale growth in a smart, predictable way and you will be giving your company a huge competitive advantage.