Businesses are very process driven. There are processes for sales, support, accounting, manufacturing, hiring, training, and virtually everything a business does. Even in our personal lives, we use processes constantly. Although we don’t give much thought to it, most of us probably shower the same way: lathering, rinsing and toweling off using the same sequence day after day. From the smallest tasks to the largest projects, processes are the way we get things done.

Processes help us work efficiently and effectively. When we realize that a certain task or activity has an element of repetition to it, and when we’re trying to achieve consistent outcomes, we develop a process for it. Once a process that works is identified, we stick with it, making improvements as we use it. In some companies, these processes are quite formalized and intentionally managed to optimize performance. There may be designated process owners who constantly evaluate the performance of their processes, looking for ways to rid them of inefficiencies and help them perform better. In other cases, the processes are informal and no one pays much attention to them as long as they work. Either way, processes are how things get done inside organizations.

Marketing is a macro process or system, consisting of many subprocesses. This view might seem to make marketing sound too machine-like, leaving creativity out of the picture. That’s not the intent of putting marketing in the context of process: the creative process is vital to marketing’s success, and it is capable of measurement, management, and improvement.

So why all this talk about processes? Processes require management to perform at peak effectiveness. You can’t manage a process that you can’t measure, so measurements—analytics—are essentially to managing marketing and all the processes of which it consists. Marketing analytics itself is a core or foundational process within marketing. Let’s revisit the definition of marketing analytics first shared in Chapter 1:

The marketing analytics process should serve as the performance interface to all other marketing processes. For everything that marketing does—email marketing, content marketing, lead generation, events marketing, pay-per-click advertising—there are ideally analytics associated with the processes. The marketing analytics process calls for identifying meaningful performance metrics for all marketing processes, tracking those metrics, analyzing them, and taking improvement actions based on the analysis.

The analytics process sounds simple enough, but it is often grand in scope considering all the things that marketing does. Consider just some of the metrics associated with the email marketing process: open rate, bounce rate, click-through rate, unsubscribe rate, and conversion. These metrics are multiplied by the number of email messages sent by a marketing organization, and it’s easy to see how quickly the marketing team is swimming in data. For this reason, the marketing analytics process is not incidental to the work of marketing, but integral to it. The dashboard metaphor for analytics is accurate: analytics provide the performance information for marketing leaders to drive the marketing machine at peak efficiency, producing the best possible results.

There’s an old joke about the instructions on a bottle of shampoo: lather, rinse, and repeat. If you follow those instructions literally, you would find yourself in an infinite process loop, until you ran out of shampoo. When it comes to marketing analytics, however, we want to interpret those instructions literally, making a continuous cycle of measure, analyze, and improve. There’s limited value in going through the process cycle only one time. Marketing analytics is most effective when there is the discipline to repeat the cycle continuously. The balance of this chapter will review the analytics process in greater detail.

Step 1: Identify Metrics

In any endeavor, marketing or otherwise, measurements are critically important because they are the means of judging success or failure. Marketing is often considered an artistic endeavor, and as such, many don’t believe its true impact is measurable. There’s no denying the artistic and creative aspect of marketing. In fact, great creative talent and work is a major differentiator between merely good and truly great marketing efforts. But just because marketing is one of the most creative business disciplines should not exempt it from measurement. In this modern era, the debate about measuring marketing’s results has been settled in the affirmative.

The fact is that in most organizations, marketing’s raison d’être is to produce revenue, either directly or indirectly. Although it is true that marketing has other missions, such as serving as guardian of the brand, if it is not a key variable in the revenue equation, most CEOs will dismiss marketing as a failure. So measurement is very important. The first challenge of measuring marketing is determining which metrics to use. This is no small matter, for the success of the entire analytics process hinges on the proper selection of metrics.

In this discussion of marketing metrics, it’s helpful to consider sales, a part of the organization with a long history of being metrics-driven. The relationship between marketing and sales is close one from a mission and goals perspective. Both functions are tasked with generating revenue, but they go about it very differently. The measurement of sales results is usually pretty straightforward. Revenue goals are set, quotas assigned, and sales are tracked, quite often to the individual representative responsible. It’s not difficult, particularly in B2B sales organizations, to know who has contributed the most to attaining revenue. The relationship between sales activity and revenue is very direct and thus fairly easy to measure.

It’s not always this clear in marketing. The reason isn’t because marketing has no effect on revenue but instead because the path to revenue for marketing’s activities is more indirect. It’s simply harder to connect the dots between the work marketing does and the results it produces. This indirect relationship between marketing’s output and revenue creates challenges to identify meaningful, relevant metrics. It’s not an impossible task, but it requires careful deliberation and selection of metrics.

One of the challenges in identifying the best set of metrics for the marketing analytics process is the sheer number of metrics linked to the work of marketing. The average marketing organization can actually produce volumes of data about most of its work, but not all of that data provides insights into real results, nor does it have meaning outside of the marketing organization.

If numbers are what the organization wants, marketing can provide them. Hundreds of metrics are available. Social media gives us likes, shares, posts, tweets, and other metrics. Email marketing generates opens, bounces, click-throughs, unsubscribes, and more. Websites let us track visits, unique visits, referral sources, search terms, and much more. There are conversion rates, content downloads, views, and a seemingly endless stream of metrics at the disposal of the modern marketing team.

Each metric has potential value in helping marketing derive its revenue impact, but almost none of them are a direct tie to a revenue figure. It’s certainly expedient to simply track data that is easy to collect and then report it, particularly when there is pressure to show some accountability. It’s unwise, however, for marketing to simply shove its full collection of metrics at the organization to satisfy the need for accountability. Although the myriad available metrics show evidence of activity, they don’t really answer the question about how marketing is performing. Marketing must be selective about which metrics it chooses, sticking with the most relevant ones so that viewing a portfolio of metrics doesn’t feel like drinking from a fire hose.

Another challenge to selecting metrics for the analytics process is that so many of them don’t have much meaning outside of the marketing department. This doesn’t mean they are useless, but it does infer that if marketing chooses to use metrics not well understood by nonmarketers, there is an obligation to explain what they mean and why they’re important.

For example, consider an impression. Marketers understand that impressions are simply a count or measure of exposure to some form of content, traditional or digital. Measuring impressions is a pretty common metric for many things marketing does, such as various forms of advertising. Marketing is always interested in garnering as many impressions as possible, and digital impressions are fairly easy to track. Imagine this scenario: a CMO gets budget approval to run a digital marketing campaign. The campaign performs well, compared to similar, previous campaigns, producing 15 percent more impressions than expected. The marketing team is elated at this result.

Should the CFO ask how the campaign went, it’s very easy for the marketing team to report the impressions data. But not all CFOs or CEOs understand exactly what an impression is, nor do they understand that not all impressions are of equal value. Furthermore, they’re not terribly interested in impressions anyway: they want to know what those impressions translate to in terms of revenue.

Impressions data from a campaign can help the marketing team improve the performance of existing and future campaigns. If the performance is low, marketing should investigate why; usually the problem is that the messaging is wrong, the medium is wrong, or both. Analyzing the data can help the team improve campaign performance or replicate great performance. As H. James Harrington said, “Measurement is the first step that leads to control and eventually improvement.”¹

As the campaign impressions example illustrates, impressions aren’t useless as a metric, but they don’t tell the story the C-suite longs to hear and marketing should want to know: the revenue impact of the campaign. To be fair, most marketers understand the need to translate metrics like impressions, visits, downloads, and so on into revenue. The problem is that they often don’t have the systems to collect the necessary data to do this with precision. Of course marketing can always estimate, taking the actual number of impressions and applying some historical, believed-to-be-accurate conversion rate and derive a figure of revenue generated. But as soon as they does this, it has left the realm of certainty and entered the realm of speculation, and it exposes itself to credibility issues as a result.

A final challenge worth discussing is the difficulty of finding metrics that measure things that are important. For example, take the concept of brand equity, an asset that is grown through marketing. By definition, brand equity is the mindshare and recognition that a brand enjoys and the benefits that come with that. Brands with a lot of equity can charge premium prices, enjoy a more loyal following, and have a degree of immunity from competitive threats. Yet brand equity is difficult to measure. Organizations that measure this often use some combination of metrics, such as market share, unaided recall, Net Promoter Score, and referrals. There is rarely a dispute about the importance of growing brand equity, but measuring it is another matter entirely. As a result, it makes the business case for campaigns designed solely to increase brand equity more difficult to prepare and justify.

These challenges—the sheer number of potential metrics, the fact that many of them don’t relate directly to revenue, and the reality that some important things are difficult to measure—seem to imply that identifying metrics is like navigating a minefield. It doesn’t have to be like this. When determining how to best measure the results of marketing’s efforts, follow these guidelines:

• Focus on objectives
• Measure efficiency
• Measure effectiveness

Marketing’s set of metrics must measure progress toward objectives, or they are measuring the wrong things. The ideal source of these objectives is the marketing strategy, and proper objectives are always time-bound and measurable. This strategy and supporting objectives are hopefully aligned with the company’s vision and business objectives. If this is the case, then identifying metrics is done by determining how to measure progress toward those objectives. When there is a marketing strategy with defined objectives, the objectives make selecting the right metrics pretty straightforward.

To illustrate this concept, consider as an example a very common business objective: to maximize profits. The marketing team develops a strategy to help maximize profits and develops two objectives:

1. Increase market penetration of current products in current markets.
2. Improve customer retention.

The marketing team plans and executes a series of campaigns and activities to achieve these objectives. Because both of these goals are measurable, marketing should have no difficulty understanding if the things it is doing are “moving the needle” in the right direction. For the first objective, marketing could measure market share, new customer acquisitions, and other things to know what progress it’s making. Likewise, marketing can measure customer retention, satisfaction, referrals, reorders, and other metrics to understand retention and the impact it has on revenue. It isn’t difficult to understand the revenue impact of what marketing is doing in either of these areas.

Identifying meaningful metrics is very difficult if there is no marketing strategy or set of related objectives. In the absence of a marketing strategy, where do the business and marketing objectives come from? While it is possible to arbitrarily choose some objectives to pursue, and then identify some metrics that go with them, that isn’t really a strategic approach to marketing. There is a symbiotic relationship between the marketing strategy (and its objectives) and marketing analytics. The marketing strategy should preexist the analytics process, and the objectives of that strategy should serve as the basis for identifying metrics for the analytics process.

So far, this chapter has driven home the need to have metrics that relate to marketing’s revenue performance. What’s also necessary are some metrics about efficiency that show how well the marketing team is getting its work done. These are essentially productivity measures, and most of the time, they’re for internal use, helping the CMO keep the marketing engine humming. Marketing should have some efficiency metrics that help it understand how well it is getting work done, as long as they are not the only type of metric the analytics process uses.

It’s very easy for efficiency metrics to become “vanity” metrics, because they help marketing feel good about how much work is getting done. What these metrics don’t do is tell you whether you’re working on the right things. When you’re using efficiency metrics, it’s important to remember that it is possible that it could show that you’re doing a wrong thing very well.

Because efficiency metrics only tell marketers if they’re doing things right, another type of metric is needed that helps marketers know if they’re doing the right things. These are effectiveness metrics, and they help marketing measure its impact. Examples of effectiveness metrics would include conversion rates, a behavioral indicator that means a prospect voluntarily took the desired action that moves them further down the sales funnel. Although this and similar metrics don’t always mean a sale is imminent, it’s a step toward a possible sale. It’s far easier to connect effectiveness metrics to revenue.

Identifying metrics is the critical first step in the marketing analytics process. Without the proper metrics, the remaining steps in this process are of little value. Generally, marketers that are new to analytics will identify a core set of metrics that are refined over time and expanded through experience.

Step 2: Analyze the Metrics

The second step in the analytics process is analysis: taking the data and, through inspection and analysis, turning it into actionable information. The goal of this analysis step is to draw insights from the data about marketing’s progress toward achieving its objectives. These insights aren’t always obvious, so analysis and interpretation helps them come to light.

During the first step, we identified the metrics that are the crucial inputs to the marketing analytics process. With the proper metrics identified, the analysis step begins by capturing and tracking these metrics. In this modern era, where almost all marketing is digital, this means having the right systems and mechanisms are in place for reliably tracking metrics and making them available for analysis.

The place where many marketing organizations were first introduced to analytics is through their websites. The website is one of the most important marketing channels for a company and capturing analytics data about website performance is imperative. Even for bricks-and-mortar businesses, a website is often the first stop for consumers who are looking for information as they consider a purchase. Every website should have analytics enabled for it so that marketers can understand performance and optimize the company’s web presence. Google Analytics is the ubiquitous, free solution most marketers use for this purpose. The realm of marketing analytics, however, extends far beyond web metrics.

If the right systems are in place to capture metrics, understanding the current state of things is not difficult. The modern marketing organization uses three core systems to capture the analytics data it needs: web analytics, CRM, and marketing automation. There are, of course, other systems that provide data to marketing as well, such as purchase transaction data, but these three systems are the pillars. Without the data collection mechanisms these key systems provide, an analytics process is dead in the water. There is simply too much going on in the modern marketing organization to track metrics data without the benefit of automation and technology.

With the right systems in place, a marketer can quickly get a snapshot of current performance metrics. In just a few clicks, one can see how many unique visitors have hit a campaign landing page, how many emails were opened, how many clicks a pay-per-click (PPC) scheme generated and more. Most marketing organizations organize their metrics into some sort of dashboard to visually present the data, making it easier to communicate and understand. Chapter 7 discusses the use of dashboards in the marketing analytics process in greater detail.

When analyzing marketing metrics, what marketers should do is understand the current state, compare it to the ideal state, and then do root-cause determination to explain any differences.

Seeing the current state of marketing’s performance by reviewing the latest analytics data is interesting for sure, but it’s just the beginning of the analysis process. The first question that reviewing this data prompts is usually: are these results good? Knowing whether any particular measurement result is good, bad, or indifferent requires some sort of comparison, to historical data or a set of benchmarks that represent a standard of excellence, preferably the latter.

Comparison is important because without it, the metrics provide no sense of direction. Consider the very important metric of customer satisfaction. An initial, quantitative assessment of customer satisfaction has limited usefulness. To illustrate this, suppose a survey of customer satisfaction is conducted using a scale from 1 to 5 where 1 equals very satisfied, and 5 equals very dissatisfied. This survey result reveals that the average satisfaction level is 2.3. How should the organization feel about this? This result is on the “satisfied” end of the scale, so there is some comfort in that, but this single snapshot of customer satisfaction says nothing about where this key metric is going.

What’s ideal is to have the data from the previous measurement period for comparison; even more useful is to have historical data to understand the trend over time. If we’re looking at landing page conversion rates, how have they performed historically? The same is true for almost any marketing metric. When there is improvement from one measurement period to the next, why did it occur? Likewise, when there is degradation from one measurement period to the next, it’s important to investigate to identify the reason. The root causes of changes in metrics are not always easy to identify. Analyzing the data is an exercise that sometimes is akin to detective work.

Benchmark data from external sources isn’t always available, but when it is, it provides a useful comparison. For example, webinar benchmark data tells us that the ratio of registrants to attendees averages 42.9 percent.² Less than half of the people who register for webinars actually attend them. With this data, companies that produce webinars can know if their registration-to-attendee ratio is cause for alarm or celebration, and they can do something about it if they don’t like how their figures compare to the benchmark.

This analysis step of the process is about drawing insights from the data about marketing’s effectiveness. Analyzing all the metrics marketing collects about its work is important because it is the way we understand marketing’s performance. It is analogous to monitoring the gauges on the dashboard of a car. When a gauge goes into the red, we understand there is a problem, but we don’t necessarily know what the problem is. The gauge gives us some idea of where to begin looking for the source of the problem. The diagnosis and prescription is the subject of the next section of this chapter.

The analysis of the marketing metrics provides a barometer of marketing’s performance. Because this analysis provides the basis for taking improvement actions, it is very important that the data used for analysis be accurate and complete. It’s possible to do excellent analysis on bad or suspect data. The marketing analytics process requires accurate, complete data, and attempting to execute the process without it will lead to erroneous conclusions and ineffective attempts to improve what marketing is doing. When this happens, as it sometimes does in the zeal to become analytics driven, it damages marketing’s reputation and credibility. Marketers must ensure that the data for the analytics process is accurate and reliable.

Step 3: Take Improvement Actions

Analyzing the marketing metrics tells us what is happening and why, but it doesn’t always tell us what to do about it if the results don’t meet expectations. Nor does the analysis always help us know how to replicate what is happening when outcomes exceed expectations. What marketing must do in this step of the analytics process is determine what actions and changes are most likely to yield improvement. No matter how thorough the analysis, it is of no value if it doesn’t lead to some sort of improvement action.

Sometimes, the improvement actions are obvious. For example, analysis of a mass email send that performs poorly reveals that an unusually high number of email addresses bounced back, indicating they were invalid. The improvement action here is obvious: remove the invalid addresses and keep the email contact list current. Improvement actions aren’t always so obvious in other cases.

In another scenario, a PPC ad campaign doesn’t perform well. Comparing this underperforming PPC campaign to previous campaigns shows the click-through rate (CTR) from the landing page is well below the norm. The analysis shows that the same strategy was in use for all PPC campaigns: a unique landing web page was set up for those who clicked on the ad. The offer or “call to action” conversion on the landing page was similar to previous campaigns that did perform well. The traffic to the landing page was within expected parameters. So what’s wrong, and more important, what is the right improvement action?

In this scenario, there’s clearly some problem with the landing page, but the analysis doesn’t tell us how to fix it; it only tells us where the problems seems to reside. In this example, it is the landing page itself, because it is not ­producing the conversions at a rate similar to previous campaigns. To find a solution, the marketing team could perform some A/B testing, a process referenced briefly in Chapter 3.

A/B testing involves presenting two versions of a landing page to visitors. The A version is the control page and the B version has some variation. The variations might include different call-to-action text, graphics, a different page layout, color scheme, or any conceivable change. The goal is to find what variations perform the best. Both pages are presented to visitors, and the analytics are monitored closely. The visitors themselves determine the winning page. In this example, the winning page is the one with the best conversion rate. It is likely that a series of A/B tests are needed to fully optimize a campaign.

A/B and other kinds of testing and research are very helpful tools in determining which improvement actions marketing should take. Testing and research help eliminate the guesswork from improvement actions, but too few marketers take the time to do it. The enemy of the testing and research approach is expediency and the mind-set of “we’ve always done it this way.” It takes time and discipline to test, research, and determine what to do as a result. Marketers under a lot of pressure to produce immediate results are often reluctant to do market research and test approaches before committing to them, as it simply takes too much time. But it’s time well spent, because it helps ensure that improvement actions are the right ones the first time. Testing and research, therefore, are excellent initial steps to identifying improvement actions.

Of all the steps in the analytics process, the improvement step is the least scripted. The range of potential improvement actions is limited only by the creativity of the team tasked with identifying them. When searching for improvement actions to take, brainstorm all possible ideas. Don’t reject any ideas in the beginning. Then, before implementing improvement actions, narrow the list of candidate ideas down to one if possible. Too many simultaneous improvements will make it difficult or impossible to know which one contributed the most to better performance.

Are improvements always necessary? Is it ever possible to skip the improvement step because things are going so well? In theory, any process always has room for improvement. In practicality, there is a point of diminishing returns on improving something marketing is doing. If there are performance objectives or standards for marketing tasks or activities, and the metrics analysis indicates that marketing consistently meets or exceeds them, then marketing can look elsewhere to invest its time and improvement efforts. Keep in mind that the scope of what marketing should measure is broad, so success in one area doesn’t mean it’s all right to ignore the entire analytics process. It just means that the area where success is occurring, based on the metrics analysis, frees marketing to investigate and improve elsewhere.

This improvement step isn’t just about addressing low performance. It’s also about replicating what’s working well. The analysis step will identify excellent performance, and when it does, the marketing team should determine how to apply what caused it to future efforts. The causes of ideal outcomes form the basis of best practices that help the organization consistently achieve the desired results.

Do It Again

The three-step process described in this chapter represents how the marketing analytics process should be executed. Each organization will have a unique variation of the process, with different metrics, different conclusions from the analysis, and different improvement actions to take. Yet the three steps remain constant: measure, analyze, and improve.

Executing this process just once or intermittently will produce some benefit, but its greatest value comes from continuous use. Repetition of the analytics process is necessary to produce consistent and sustained improvement. The real advantages of analytics only come with ongoing, long-term execution of the process. Every time marketing does work in pursuit of its objectives, that work should fall within the scope of the analytics process.

For every marketing campaign, communication, event, and investment, the cycle of measure, analyze, and improve should occur at the conclusion before the next campaign, communication, event, or investment takes place. The frequency of invoking the analytics process will of course vary depending on what marketing is doing. Some marketing actions are so important that analytics are monitored in real time, whereas others require daily, weekly, or monthly scrutiny.

What’s important is to not let complacency infiltrate the process. It’s tempting to relax when regular reviews of the metrics show that everything is fine. In fact, in these circumstances, it is very easy to assume that marketing’s work is so fine-tuned that it can run on auto-pilot. This state of marketing nirvana, is rare, and when it does materialize, it never lasts for long. Remaining vigilant to the analytics process is the way to ensure the best long-term performance of marketing.

__________________________________

¹Joseph L. Levy, “In My Opinion,” CIO Enterprise, September 15, 1999, p. 10.

²ON24 Webinar Benchmarks Report, 2014 edition. http://www.on24.com/wp-content/uploads/2014/04/ON24_Benchmark_2014_Final.pdf.