Measuring the effectiveness and return on investment in online marketing

Knowledge in performance measurement of online marketing is still not exactly at its peak. Everybody wants to invest without any risk and for every invested euro expects two euros in return.

Let’s face it. The investment without risk does not exist! Even the most reliable investments can backfire. In a complex digital world the success depends on a great number of variables. However, this does not mean that the risk can’t be minimized. This is definitely possible, but it requires a good understanding of online metrics.

Overall situation in this area is pretty bad. Not only that minimizing the risk in online investment is more an exception than a rule, but also most of the approaches to the measurement of online investments are incorrect. In this article I wanted to clarify a few basic rules for measuring online investment and furthermore upgrade it into a more complex and integrated view. The topic is definitely too broad to sum it up in one article, that is why some of the models and explanations are generalized.

For better understanding I am using an example of online advertising with Google AdWords throughout the article. Our imaginary advertiser has a budget of 1.000 € for monthly campaigns in online advertising, wherein the amount is divided into the Search and Display Network. Below is a table of the basic advertising metrics.

Minimizing the risks of an investment

If we want to minimize the risk of investment, we need to measure and regularly (in reasonable intervals) monitor return on investment.

To calculate ROI, the benefit (revenue) of an investment is divided by the cost of the investment; the result is expressed as a percentage or a ratio.

ROI = profit of the investment / cost of the investment

or

ROI = (revenue of the investment – cost of the investment) / cost of the investment

If the result is negative (return is lower than the costs), the investment does not pay off. The higher the ROI, the more profitable the investment.

The trick in online marketing is that we need to know the benefit or the result if we want to calculate the investment. Too often we are using wrong data, which have no connection with the investment. In economics, investments are usually calculated on the basis of cash. How much profit is generated by certain investment, is quite simple in financial world. But in online marketing the result of an investment depends on the purpose of the investment.

Let’s calculate ROI in our example:

ROI = (5100 € – 1000 €) / 1000 € = 4.1

The calculation shows that our investment is profitable but we have made a crucial mistake as we used the wrong data. Due to excessive focus on sales we took the revenue as a result which in the case of Display Network doesn’t make any sense because its primary aim is not to promote the purchase. More follows below.

We get the correct calculation of ROI if we choose the right metrics to measure the result. Only in rare cases we use revenue or financial benefits (€). The current situation is unfortunately totally the opposite – due to the expected immediate impact on sales, the majority of investments in online marketing is measured by euros.

When the result of the investment can’t be defined in monetary unit, the above formula for calculating the return on investment becomes more or less useless. Therefore, in the case of online investments, in most cases we use other metrics – cost per conversion or CPA (cost-per-action).

In our case, CPA (cost of purchase) is 19.60 € (1000 €/51 purchases), but we just made the same mistake as explained above and used wrong information in calculation.  

How to choose the correct metrics to calculate the cost of conversion in the online marketing?

In online marketing everything starts with setting the objectives. The objective defines the purpose of the investment and thus the metrics that are appropriate for the calculation of the cost of conversion. The objectives in online marketing do not differ significantly from conventional marketing goals. The easiest way to define them is through the consumer decision journey process. Each stage of the consumer decision journey represents dimension, inside which we can measure results by means of metrics.

The easiest way to explain this is in the phase of purchase, where the dimension is the purchase, and the metric that measures a result is the number of purchases.

Clearly defined goals in online marketing determine which phase in consumer decision journey we want to target and consequently which metrics do we want to use. Below I present a table with phases in consumer decision journey, potential goals and associated metrics.

Using the table above we can easily calculate the cost of conversion in online marketing. Conversion happens when we reach the goal of each phase in the consumer decision journey. Conversion in the awareness phase happens when we show our ad; and in the purchase phase when user buys something. The calculation of the exact number of conversions in the awareness phase is trivial, since the majority of ad networks offer this information. Calculations become more complex when we talk about other phases in consumer decision journey. 

The main point in calculating the CPA is not monetization of these values, but comparison between different investments. This approach allows us to compare the effectiveness of different marketing channels in the individual phases of the consumer decision journey.

The table also shows us the right direction when selecting the combination of goals, conversions and metrics. Combining the goals from first phase and metrics from the second phase leads to wrong assumptions and conclusions, and consequently to incorrect calculations of the conversion and return on investments (as shown in the above examples). 

Let’s take a look how to calculate the actual cost of conversion in our case. Due to the fact that we use two different types of advertising, we also target two completely separate phases of consumer decision journey. For easier understanding let’s assume that we are using Google Display Network to increase brand awareness (metric = the number of impressions), and Google Search Networkto increase the number of purchases (metric = number of purchases).

As mentioned above, the calculation of the cost per action has to be done for each phase separately.

Display Network (CPA = cost per 1000 impressions)
CPA = 500 € / 1250 = 0.4 €

Search Network (CPA = cost per 1 purchase)
CPA = 500 € / 50 = 10 €

We found out, that generating one online purchase costs us 10 €, and generating 1000 impressions of our brand costs us 0.4 €. Investments into these two different phases mustn’t and can’t be compared.

To be more clear, I present few examples:

• If you manage a Facebook community, ask yourself why you are doing it. Measuring the number of purchases or even »likes« is absurd. Facebook community has one and only metric, and this is engagement.
• If you advertise on Google Display Network or on some large website with an intention of increasing the brand awareness, it is pointless to measure the purchases or even user engagement. Key metric in this case is the relevant number of ad impressions. It is wise to take into account the cap limit of ad impressions per person.

For basic calculation and comparison between digital channels this will be enough. However, the whole story does not end here. As I mentioned in the introduction, the success of a project in a complex digital world is affected by large number of variables. Although it seems that calculating the cost of conversion cost is simple it actually isn’t.

How to calculate the actual cost of conversion?

Consumer decision journey process is not linear, but extremely intertwined; it goes through different digital channels and often extends to offline world. The above mentioned factors have a significant impact on the number of micro conversions in each phase of consumer decision journey and without taking them into account we limit ourselves to a narrow number of »last click« conversions (conversions, which are attributed only to last user interaction with the brand).

Online 2 Offline

Ignoring the comparison between different digital channels we will focus now solely on the online – offline relationship.

In the phase of preference we come to the dilemma of how to accurately calculate the number of micro conversions or user engagement with the brand. Engagement can be a comment or Facebook like on a blog, a subscription to e-newspaper, download of the brochure, and contact via e-mail or phone. Almost all of these forms of brand consumers’ involvement are measurable using web analytics except the last two. To calculate the exact number of conversions in the preference phase we must monitor the number of contacts with customers outside of the website. This can be done by using different email addresses and telephone numbers.

The percentage of users who will as a result of your online campaign get in touch with you in offline world is around 50 %. This percentage is getting higher towards the end of consumer decision journey.

The actual number of conversions in preference phase = number of online conversions * 2

According to the Consumer barometer 65% of Europeans do a research of their purchases online. Despite of an online research, 2/3 of them make a purchase offline. The data show that in Europe, as a result of consumers’ online research, on every online purchase there are 2 offline purchases. Therefore, in order to calculate the correct number of purchases as a result of online marketing, we have to take a number of online purchases and multiply it by a factor of 2.82.

Actual number of conversions in purchase phase = number of online conversions * 2.82

In a loyalty phase this factor is in most cases even higher, as consumers become accustomed to the brand and develop a personal relationship with it. Probability to come into physical contact with it is even higher. A great help can be a calculation of the lifetime value of the customer.

This approach allows us to accurately calculate the number of conversions which are a result of online activities and with this we can effectively compare online marketing (as a whole) with other forms of marketing. A quick calculation shows that the classical approach to the calculation of the number and cost of conversion mentioned at the beginning of the article considerably underestimates the value of online marketing.

In our imaginary case we should also consider the factor 2.82 when calculating the number of conversions in purchase phase. Actual number of conversions in purchase phase (derived from advertising in Display Network) is 141, which drastically reduces the cost of conversion, from initial 10 € to 3.54 €.

A comparison of the profitability of different online channels

Calculating the cost of conversion in each phase of consumer decision journey is simple. It is more difficult to compare the costs of conversions for each online channel. When measuring the results of each online channel we stumble upon a lot of things.

As already mentioned, the conversion rarely occurs as a result of the linear path through consumer decision journey. In order to achieve conversions we use several different channels. The first problem arises in the process of allocating the value of conversion of the different channels, involved in the conversion. To give all the prominence just to the last channel (last click conversion) would probably not be the most fair.

Additional problem is use of different devices. We mostly use around three – laptop, desktop, smart phone, tablet – and monitoring users through all of them is/was virtually impossible.

We are facing those challenges across all digital channels, that is why we can neglect mentioned factors (of course, we assume that we have used certain digital channel in the right phase of consumer decision journey). Factors are eligible only if we want to monetize value of certain digital channel.

Monetization value of each digital channel

Monetization value of each digital channel is particularly important when we want to justify the investment in specific channel. Based on the monetized value and according to the formula, given at the beginning of the article, we can calculate the real return on investment.

Monetization of each channel begins by calculating the number of macro conversions (e. g. purchases), which we can actually evaluate with the currency (€).

1. As a starting point we take the number of purchases, provided by the chosen channel on the basis of last click attribution – thus directly online.
2. To this we add share of purchases, where the selected channel was involved in the purchase process (attribution).
3. The sum is multiplied by a factor of usage of different devices and by a factor of users who do not use cookies.
4. In this way we get an estimated number of online purchases where the selected channel made a contribution. This number is multiplied with Online 2 Offline factor in order to get the number of all purchases.
5. In the end, we add the factor of the lifetime value of the consumer.

With this we calculated the actual number of macro conversions credited to the selected channel. To calculate its value, we take the actual number and multiply it with an average value of the conversion. The result is the value of each digital channel from which we can calculate its profitability. Usually, the actual profitability of digital channels is 3 to 4 times higher than baseline.

Value of digital channel = actual number of conversions * average value of conversions

Let’s repeat the whole process to calculate the value of Google AdWords advertising :

1. As a starting point we take the number of purchases, provided by the chosen channel on the basis of last click attribution – thus directly online. In this case we consider all conversions (Search and Display Network) = 51.
2. To this we add share of purchases, where the selected channel was involved in the purchase process. We get this data in Google Analytics. In this case we estimate that the number is 6.
3. The sum is multiplied by a factor of using different devices and by a factor of users who do not use cookies. An estimation for European average is3 %. Factor is therefore 1.03.
4. In this way we get an estimated number of online purchases where the selected channel made a contribution. This number is multiplied with Online 2 Offline factor in order to get the number of all purchases. This factor is 2.82.
5. In the end, we add the factor of the lifetime value of the consumer. In this case, we take into account that each newly acquired customer did another 5 purchases, the percentage of newly acquired customers is 50%. Thus, by using the attribution model we assign an additional 10% of value to the first purchase.

The number of macro conversions, credited to Google AdWords in our case, is:

(51 + 6) * 1.03 * 2.82 * 1.10 = 182

If we consider that the average value of the purchase is 100 €, the value of digital channel is approximately 18.200 € and ROI is 17.2. Quite different from the initial numbers, right?

Consistency of digital channels with the objectives of online marketing

More than mere monetization of certain digital channels, it is important to know what each channel contributes to the total value of digital marketing and whether each digital channel is used in the right phase of consumer decision journey. In particular, we have to be extremely careful with the latter.

As mentioned at the beginning of the article, the target defines the purpose of the investment, and thus also the metrics. On this basis, we choose a suitable digital channel. If we set the objective to increase brand preference (metric = engagement), then achieving results f (e. g. a large number of ad impressions) from other phases of consumer decision journey does not bring desired results. In that case, we should not fool ourselves, but we must act quickly in accordance to achieve them.

Often a particular digital channel does not bring the desired results, but this does not mean that digital channel “does not work”. Maybe it was only used for the wrong purposes. Fortunately, Google Analytics offers a tool with which we can analyze how different digital channels coincide with the different phases of the consumer decision journey. The tool is called “Multi-Channel Funnels” and is located in the “Conversions”.

With a little knowledge we can determine how users on our site convert and in what phase of consumer decision journey they use certain digital channels.

In our case, just a quick look at the initial metrics table shows us that the use of the Display Network in the purchase phase does not make any sense, since the key performance indicators are significantly lower than those in the Search network.

Once you know the real purpose of the use of certain digital channels, monetization of its value is no longer meaningful. More crucial becomes the process of evaluating each phase of the consumer decision journey.

The evaluation of each stage of the consumer decision journey

Despite the fact that the number of purchases is the most important metric, the entire marketing budget is never devoted solely to this phase. Each stage of the consumer decision journey contributes its share to the final success. The trick is to figure out what is this share.

For an optimal allocation of resources across the entire consumer decision journey it is therefore necessary to know the contribution of each phase to the final success. There is no uniformed formula to calculate the contribution of individual phases, because each brand is in a unique position, influenced by the endless number of factors such as market share, brand awareness, trust and reputation, user experience, price, competition, seasonal trends, macroeconomic factors etc.

The evaluation of each stage of the consumer decision journey is a matter of testing different attribution models, among which we can take as a base the linear attribution (evenly distributed investment to all phases) or time decay attribution (contribution is decreasing from purchase phase towards the awareness phase). In case you already know your marketing mix, you can begin to use the more advanced models. Important role in building a good attribution model has also the knowledge of conversion and loss of users through each phase of the consumer decision journey.

As mentioned above, there is no single attribution model that would work for all. It is necessary to test various models and hypothesis. Fortunately, digital marketing with its advanced analytical tools allows us to do that.

Optimization of online marketing

Good understanding of web metrics and ways of measuring performance not only allows us a better understanding of online marketing and more transparent investments in it. Based on the right set of approaches to calculate the performance of online marketing we can optimize our long-term web presence and thereby dramatically improve its performance.

Back to the beginning

1. There are no investments without risk but with a good knowledge of online marketing, especially with an understanding of individual digital channels, we can build an optimal online performance and minimize risk.
2. If we want to assess the performance of a particular investment, we first need to know its goals and select the correct metrics, leading to the result. Mistake in this step can be critical.
3. We need to bear in mind that the internet does not work as isolated marketing channel but connects various digital and offline channels which form a complex marketing mix. To calculate the value of online marketing, we must also take into account its impact on other marketing channels.
4. The real value of each digital channel can be calculated only when considering its impact on other digital channels, technological limitations in monitoring the conversions and its impact on offline channels.
5. The correct approach to digital marketing will be achieved only when we are fully aware of our marketing mix and when we place digital channels in the right phase of consumer decision journey.
6. Optimizing digital marketing will be achieved only with a good understanding of attribution model of each phase in consumer decision journey.