A Quick Glance at Complex Marketing Metrics (feat. LTV)
The Cs, the Rs, Another R, and the Us
DelightRoom operates Alarmy, a premium alarm app that surpassed 49 million cumulative downloads worldwide as of September 2020. With just over a dozen members, our structure is already too busy keeping up with the day-to-day work, but exactly because of that, we voluntarily run internal study groups to broaden our thinking and handle our respective tasks in a better way.
The current study groups include 'Dev Study', where Android and iOS engineers gather to read and debate new development books, and 'DDL (Dany's Data Lab)', spearheaded by engineers but joined by members from diverse backgrounds to share insights and deeply apply various features of the user behavior data analysis tool Amplitude based on those insights.
What I'll introduce today is an excerpt from a marketing metrics sharing session requested by the presenter of the DDL study group. If you're not in marketing, you might partially know some marketing metrics depending on your collaboration touchpoints with the marketing department, but you've probably felt confused on how to understand seemingly similar yet different metrics when they suddenly pop up during a meeting or work. As with all tools, knowing what they are useful for and actually utilizing them is much more important and meaningful than memorizing their names and rattling them off. Thus, I prepared this sharing session with the following goals.
1. By understanding the structure and principles of marketing terminology, enable inferring the meaning of any unknown metrics that appear in the future.
2. Understand the correlation between metrics and utilize them meaningfully in the decision-making process.
Even if you're not a marketer, understanding frequently used marketing metrics structurally helps you gain insight into users (consumers), the service (product), and the business. So let's keep these four properties in mind. They are the cost-related metrics Cs represented in currency, the revenue-related metrics Rs, another set of metric Rs expressed as ratios, and the user-related metrics Us indicated by general numbers. Just knowing C, R, R, U allows you to roughly categorize most marketing metrics.
If you're at a company operating an app, a web service, or marketing utilizing an app or web, you'll frequently encounter the metrics below even if you're not a marketer. You could immediately categorize them using the four properties above and figure out their meanings, but as humans with an instinct for exploration, you might wonder why on earth so many metrics were created.
Reason why?
Why exactly were all these complex marketing metrics created? Can't we just look solely at 'revenue'?
To fundamentally understand this, let's lightly touch upon the theoretical basics. Looking back at the dictionary definition of marketing makes it a bit clearer.
What is Marketing?
Systematic management activities to efficiently provide products or services to consumers.
This includes market research, commercialization planning, advertising, and sales, aiming to give consumers maximum satisfaction and most efficiently achieve the producer's production purpose.
In other words, marketing is an activity that delivers products and services at the optimal cost and maximizes the results obtained from consumers. To efficiently achieve goals according to the 'production purpose', you have to measure from various angles whether you are actually getting closer to the goal. That is why various metrics were born. In some cases, the production purpose is not revenue. Because the focused goals differ depending on what each company—be it a game company, NGO, research firm, or distribution company—pursues, the target substituted into each metric and the name of the metric itself vary as well.
In a more subdivided concept from marketing in a broad sense, the activity of repeatedly measuring and analyzing numerical performance to maximize the desired results is called performance marketing.
Performance Marketing = Tracking & Optimization
Performance marketing is a series of processes that constantly measures marketing performance and optimizes overall marketing activities like ad costs, channels, and duration based on that data. The purpose is to find customers willing to continuously pay for our products and services and bring them in at the minimum cost.
That means we have to track and uncover many question marks: what the numbers for money and customers are and the ratio between them, how many users are brought in at what cost, and subsequently how long they stay and how much revenue they generate. To grasp how much cost per result (C) we are paying to get the desired outcome, how much average revenue (R+U) we are being paid per paying user, what percentage is the ratio (R) of the desired result occurring in an ongoing campaign, and how many users (U) have signed up for the service, we must refine the collected information so we can read it efficiently. As this necessity emerged, individual tools that allow effective measurement of the aspects of cost, revenue, ratio, and users were born, and their sum is the marketing metrics that look so complicated.
Some metrics have only one property among C, R, R, U, while others are a complex association of multiple properties. Just as a non-expert looking at a forest might not accurately distinguish if a tree is a fir or a pine but can broadly tell if it's a conifer or a broadleaf by its appearance, once you know what the four properties are, you can roughly categorize the metrics and intuitively understand them.
Then again, let's categorize the various metrics listed earlier into the four properties. Among cost-related metrics, the ones starting with CP will notably catch your eye. You can see that revenue-related metrics start with AR rather than just R, and ratio-related metrics have a mix of ones ending in R and starting with R. User-related metrics mostly end in U and look straightforward and clear compared to the others.
1. Cost-related Metrics, the Cs
Most metrics starting with C are metrics created to calculate the cost for whatever target comes inside the parentheses after Cost per. Along with the most widely used unchanging metrics—CPI (Cost Per Install), CPC (Cost Per Click), and CPM (Cost Per Mille / 1000 impressions)—CPA is the most representative one, whose meaning can change depending on what the concept of 'Action' is set as. Usually, CPA is heavily used to calculate the cost for a specific 'action' that passes the first gateway to becoming a real user beyond just a download, such as signing up, registering, applying for a free trial, or participating in an event. CPE can represent the cost per engagement like a like or share, or it can mean the cost per 'execution after install', which is a heavier step than CPI but lighter than CPA. Many app marketing agencies and ad network companies use this CPE as their billing standard.
So, what is CAC, starting with a blue C that has been bothering you at the very top since earlier? The C in CAC stands for Customer, meaning user. Customer Acquisition Cost is literally the cost of acquiring a customer. It's all the costs incurred until finding 1 true customer who has paid for our service or product. You could say it's the final boss encompassing the CP- metrics. If a profit-seeking company passes the initial investment stage and its business lands on a normal track to some extent, it must be able to calculate and track this CAC, and appropriately maintain or reduce this cost. Even if it's not in an aggressive marketing phase based on investment concepts, if the CAC is higher than the LTV (Customer Lifetime Value), it will be hard to sustain the business. LTV is truly important, so let's deal with it again at the very end.
2. Revenue-related Metrics, the Rs
The revenue-related metrics Rs all start with AR (Average Revenue). Where does revenue come from? That's right. As you might guess, it's the user. The three metrics above each calculate the average revenue generated from 1 'certain' user. They are the Average Revenue Per User (ARPU) for total users, Average Revenue Per Paying User (ARPPU), and Average Revenue Per Daily Active User (ARPDAU). This complex metric, combining user and revenue properties, plays a decisive role in calculating the aforementioned LTV. If the revenue generated per user shows a relatively stable trend, you can remain unswayed by the ebb and flow of user influx and churn, and predict it to convert into assisting energy to prepare for the future.
3. Ratio Indicator Metrics, the Rs
If the previously introduced Cs and Rs are metrics that split the targets of cost and revenue to show them, these Rs are metrics categorized by the notation format rather than the characteristics of the target itself. The metrics with R coming at the very end mean the ratio to the target in the preceding parentheses, and metrics starting with RO indicate the ratio of how much was returned compared to the invested cost. Strictly speaking, it seems they would belong to Revenue, but what's different is that they are expressed as a ratio, not a currency.
Among them, ROI is an economic term widely used from the past rather than a general marketing metric, meaning Return On Investment. You divide a company's net profit (revenue - cost) by the total investment amount (total assets invested in the business) and use it as a standard to judge whether to continue investing in this business in the future. Management evaluation metrics like ROA (Return On Assets) and ROE (Return On Equity) are also used to judge a company's investment value by measuring the 'Return On'. If you're interested in investing, you'll be able to encounter them often in the future.
If I had to pick the two metrics most closely touching field operations among the many Rs, they would be ROAS (Return On Ad Spend) and CVR (Conversion Rate). ROAS is simply calculated as total revenue / ad spend, or more conservatively calculated as 'revenue increase due to ads' / ad spend. If you are continuously measuring and managing LTV while ad expenditure is only occurring through app install ads, you can divide LTV by CPI to set it as ROAS and intuitively measure the performance by media channel you are executing on.
CVR can be used in various meanings like CPA depending on what the desired 'conversion' is. Sometimes, when looking at reports automatically generated by each media channel, even though there's a separate CTR indicating the click-through rate, clicks are often counted as 'conversions' and shown as CVR. If the 'most desired outcome' you want to get from an ongoing campaign is an install, it would be good to measure the CVR based on installs.
4. User-related Metrics, the Us
We've reached the long-awaited user-related metrics. Even if you're not a marketer, if you're a member of a company providing an app service, you can't help but encounter metrics that analyze users. Rather than just the cumulative number of downloads, you must know the current status of how many active users are using our service and how many people are satisfied enough with our service to pay for it, in order to proceed with work aligned with the direction. DAU represents Daily Active Users, and MAU represents Monthly Active Users. In Alarmy's case, we currently maintain over 4 million MAU and 2 million DAU.
But among the highlighted yellow key metrics, you might see a familiar blue C. Do you remember CAC, which occupied the top of the cost-related metrics? The C in CLV also stands for user, that is, 'Customer'. CLV is Customer Lifetime Value, an economic term that has been used since before app marketing existed in the world. Business owners have long realized through experience and studied that retaining regular customers at a restaurant is more efficient than attracting new ones. This CLV transformed into LTV in modern app marketing, becoming a metric that more intuitively chases the Lifetime. NRU and PU, which indicate new registered users and paying users, are also some of the widely used user-related metrics.
5. Other Major Complex Metrics
Along with LTV, Stickiness—also known as stickiness or affection level—and Retention are also among the important metrics that allow us to comprehensively judge where our app service is heading. If the ratio of daily active users to monthly active users is high, it means our service has seeped deeper into users' daily lives. In gaming terms, it means it's a fascinating masterpiece that users log into almost every day all month long, and if it's a utility app, it means it has already become a part of the user's life. In Alarmy's case, it consistently maintains a Stickiness of around 48%. The ideal value of stickiness can vary depending on the characteristics of the industry and main target, but it's important to watch its trend. If stickiness suddenly drops significantly, you need to figure out the cause and take action. There might have been a product change that lowered usability, or a campaign might have run through a media channel where the retention of new users is remarkably low. If you extremely focus on the single metric of CPI without checking how well newly acquired users are being retained, it can truly end up being like pouring water into a bottomless pitcher.
6. Again, LTV (Customer Lifetime Value)
Earlier, while explaining CAC, the customer acquisition cost, I mentioned that even if it's not an aggressive investment stage, if the CAC is higher than LTV, the business cannot be sustained. It is self-evident that if you keep spending marketing expenses at a cost that exceeds not just actual revenue but even expected revenue, you will go bankrupt. If you look at famous brands that vanished into history, a decisive factor in their defeat was continuously executing high-unit-price ads like TV CFs and PPLs even in situations where it was difficult to recover marketing costs due to poor product quality. They utilized media channels that significantly raised the CAC when the LTV was remarkably low due to a failure in user retention. Let's briefly go over the exact meaning and calculation method of LTV.
LTV: The total cost a customer has spent (or is expected to spend) on our service over their entire lifetime until they completely churn
= ARPDAU x Average Usage Period
= Specific Period Cohort Revenue / Specific Period Cohort Total User Count
= ARPU x 1/Churn (1-Retention)
The method for calculating LTV is applied differently depending on the characteristics of the company's product or service. Even with the commonly used method of multiplying ARPDAU, the average revenue per daily active user, by the user's expected average usage period, you must judge how far into the future to predict according to your needs based on the service's characteristics. Real data is needed for prediction, and a widely used method is drawing a trendline with retention data of 30–90 days or more, and substituting the future period you want to find into the derived formula to infer it. However, even if it's based on real data, a predicted value can't always hit the mark exactly. That is why a plan that accounts for risks is necessary.
7. How to Increase LTV?
If you're a business entity, the fact that you must increase customer lifetime value while simultaneously lowering customer acquisition cost is like a single grand premise. Then how on earth does LTV increase?
Looking at the three elements that make up LTV reveals the answer. Because LTV consists of monetization, retention, and virality, you can seek improvements in these three aspects.
Improvement of revenue structure through price hikes, package additions, etc.
Improvement of retention through strengthening product quality
Improvement of Organic influx rate through voluntary shares, invites, etc. not included in CAC
For example, if the monetization structure consists of just a single paid product, diversifying this to improve the revenue structure can be a huge help. You find the threshold users are willing to pay through market research, and then slightly raise the price of existing products or launch entirely new product packages that meet different needs. Even when providing the same service, you can create distinct packages by varying the scope of usable features or the usage period.
If you're in a situation where you can't touch product pricing, you can consider a direction of increasing the user's retention rate by improving the product's usability or heightening its appeal, but because retention takes time to be properly measured and has many variables, it's not an item that can deliver visible changes in the short term. However, if members act while knowing there is room for such improvement, action items will be selected much faster and specific improvements can be achieved.
Improvements in terms of virality can actually cause adverse effects if the business entity's intentions are projected too much, so rather than running blatantly obvious viral marketing, it's better to deploy sources where voluntary sharing can happen more easily. You can attempt things like planning and distributing interesting content and creating a highly searchable environment, or deploying in-app features that are easy to share. It goes hand in hand with tasks that increase organic user influx, such as Search Engine Optimization (SEO) and App Store Optimization (ASO). Organic means natural influx, and its counter concept, 'non-organic' users, refers to users brought in through ads.
So far, we've looked at the C, R, R, U classification method to see complex marketing metrics at a glance, major complex metrics, and LTV among them. No matter how complicated an object looks to use, in the end, it's merely a tool made for our convenience. I hope you all arm yourselves with today's learnings to build a better service than yesterday.
We are DelightRoom, providing the Alarmy app that wakes up 49 million mornings.