Digital Marketing Measurement

Pricing models (CPM, CPC, Flat-rate) and CTR

Display advertising campaigns are traditionally charged on a CPM (Cost Per Mille, also called Cost Per Thousand) pricing model, for example: 4 € every 1000 ad impressions. Let’s assume that my ad is delivered 100,000 times (100,000 ad impressions) with a CPM of 4 €, that will cost me 400 €.

So the formula is:

CPM = (Impressions/1000) x € rate

In the example above:

(100,000 impressions /1000) *4 = 400

Another common pricing model is CPC (Cost Per Click, also called Pay Per Click), which is what Google AdWords is based on. In a CPC ad campaign, advertisers pay only when/if users click on their ads and not based on impressions. In other words, with a CPC model impressions are free.

Another pricing model is Flat-rate, which means the advertiser pays a fixed agreed fee, usually for a defined period of time, regardless of performance factors such as volume of impressions or clicks.

Which is the best model? Obviously it depends. In very general terms, here’s a quick comparison:

From a publisher (seller) point of view:

CPM is based on traffic, so it’s a good fit for popular sites with high traffic stats: you’re selling impressions, not clicks.

Another element to consider is that regular visitors of your site – returning to your site on a regular basis, typically as direct referrals (e.g. people having your site in their bookmarks) – will be less likely to click on ads than visitors coming to your site for the first time from search engines. Therefore CPM is generally better than CPC to monetize on traffic from direct referrals.

For small sites/small bloggers CPC is a better option compared to CPM. With CPC you can get good revenue even if your traffic is not tremendously high. The important bit is the placement and the implementation. Ads need to be relevant and blend well with your content, adding value to it. Good placement and implementation will make user clicks more likely.

Flat-rate is a bit of a gamble. For example you may sell an ad placement on your site for a month to an advertiser for a given price in consideration of some traffic expectations based on the past. If that specific month your traffic is higher than expected, the advertiser will have done a good deal, whereas if you have a poor month in terms of traffic, you’ll have done a good deal, because you’ll get that fixed rate anyway.

From an advertiser (buyer) perspective:

CPC usually gives more control on your spending, because you set your budget and pay only for actual clicks, whereas with CPM you may deplete your budget quickly if your ad is performing poorly.

I suppose CPM is a valid choice if the purpose of your ad campaign is not so much about getting users to click through and get to your landing page, but more generally getting your message across by being seen. For example, if the aim of your ad is letting people know that a particular movie is out (create awareness/generate business offline) rather than getting them to click though to your site to buy a ticket, your ad may serve its purpose well just by appearing and being seen on a popular and well targeted site.

Likewise, if the key factor for you is to get a placement on specific high-traffic/popular sites that allow you to get exposure to your target audience, it’s likely that those sites will sell placement on a CPM basis or via flat-rate fees.

Flat-rate, as said, is a bit of a gamble for both sides, as your expectation is set on the basis of past performances. The risk is obviously smaller if you buy ads on well-established sites with little fluctuation in traffic.

I have talked about impressions and about clicks. A key metric to measure an ad performance on a particular site is CTR (Click Through Rate), which is essentially the number of clicks on an ad divided by the number of times that ad is shown (number of impressions), expressed in percentage.

It is important to underline though that CTR in itself is not a conclusive measurement of success for an online advertising campaign. All you know from CTR is what percentage of people exposed to the ad clicks on it, what percentage of viewers is converted into “clickers”.

High CTR doesn’t always mean success though, for various reasons. Here are a few.

For example, your ad may be catchy but misleading: users may be tricked into thinking they’ll get a free phone, whereas you’re only advertising an expensive one. Or perhaps your ad may be doing a great job in sending users to your destination URL, but then your site may be very counter-intuitive and clunky to navigate or poorly designed and you may not be able to turn visitors into customers.

Less straightforwardly, your ad may be implemented deceptively on the publisher’s website  – for instance it could be too close to navigational elements or excessively blended into the content – and users might be clicking on it by mistake. You would get unintentional, uninterested visitors and convert poorly.

Another unpleasant case could be fraud clicks from competitors – clicking your ads to boost your costs – or from publishers themselves – clicking to boost their earnings – or similar scenarios.

Bottom-line is that CTR is an important indicator of an ad performance on a given site, but not necessarily a definite measure of success, as it is an intermediate metric and doesn’t guarantee return of investment.


Display advertising: display ads and rich-media ads

A few posts ago I talked about the POEM categorization that distinguishes between paid, owned and earned media, in relation to digital marketing. I mentioned that display advertising is one of the main samples of paid media types in the digital world. I’ll try here to cast an overview of display advertising.

In terms of paid online advertising, generally 3 main categories of ad types are mentioned: rich-media ads, display ads and text ads.

Text ads are typically associated with Google AdWords, Google’s main online advertising product, which – despite nowadays offering also display and rich-media formats – has traditionally built its success on contextually targeted textual ads. I’ll talk about AdWords and text ads more extensively in a later post.

Display ads – or banner ads – usually combine text with a graphic element. They can be static or animated (e.g. flash animations). Typically the way a user interacts with a display ad is basic and straightforward: when users click on a display ad, they get to a landing page on the advertiser’s site (destination URL).

There are many different banner ad formats in use, with different sizes.

To some extent ad sizes have been standardized by the Interactive Advertising Bureau and the most popular can be identified in the following formats:

–       Leaderboard: 728×90 px

–       Skyscraper: 160×600 px

–       Island or MPU (Mid Page Unit): 300×250 px

Rich media ads are instead more advanced from a design and user interaction point of view. They can combine multimedia elements such as video, sound and animation, as well as allowing much higher user interactivity.

With rich media ads you can have fancier animation effects – such as peel-back (or peel-away) ads, floating ads, expanding ads and so on – but also more complex user interaction: you can, for instance, let users play a game or input data through the ad itself.

The possibilities at disposal of graphic designers and developers’ inventiveness are vast.

Generally rich media ads are catchier and more engaging from a user perspective and tend to have higher CTR (Click Through Rate, I’ll talk about this soon in another post) than traditional banner ads.

On the flipside though, rich media ads are heavier to load and therefore require more bandwidth and might be less accessible to users with slow Internet connections; also, they’re also more likely to incur problems of browser incompatibility or not work as intended.

Another popular ad format is the homepage takeover (see same examples here), which can have very high impact but also be quite intrusive or disorientating for users.

Site section sponsorships (e.g a mini-site hosted to the advertiser on the publisher site) are another typical advertising solution, conceptually similar to one or a few pages bought by an advertiser on a newspaper or a magazine in offline advertising.

With the increasing popularity of catch-up TV, another common ad format is pre-roll and mid-roll ads (which play before or during TV programs).

Measuring audience

In my last post I gave an overview of the main key metrics in use to measure reach in online marketing, pointing out though that web traffic metrics don’t provide much insight on “who” our audience is.

When it comes to defining audience demographics, marketers need to turn somewhere else.

For example, email sign-ups and website subscriptions are among the most common tools in use to retrieve audience information such as gender, age, level of education and income.

Another commonly used source of information for audience demographics is

If we look up – one of the main property sites in Ireland – we can see for example that, according to Alexa, “relative to the general Internet population, 25-34 year olds are over-represented at”

alexa information on

Another popular resource for audience measurement is Google Display Network AdPlanner, a free tool that helps advertisers/media buyers identify appropriate ad placements for their target audience and plan their ad campaigns as well as do research and competitor analysis.

Publishers also use GDN Ad Planner to provide an overview on reach and audience and other information about their websites.

Online surveys are another common way to gather information on audience demographics. There are many tools that help create online surveys, such as or

Another option is panel-based surveys, which use small size, people-based panels to represent the behavior of the larger Internet population. It is to be taken into account that usually panel members are offered incentives to participate in the surveys, which introduces an element of bias in the sample.

Ultimately, from a publisher’s perspective, being able to provide an accurate audience profile of your website – as well as detailed information on your web traffic – is vital to attract advertisers’ interest, whereas for advertisers, identifying the most appropriate ad placements is vital for the success of an advertising campaign.

Measuring reach – Key metrics for web traffic measurement

In today’s online advertising landscape, measurement is critical for both publishers and advertisers/media buyers.

Advertisers want verifiable, comparable metrics to guide their advertising campaigns, in order to identify the web content that best allows them to reach their target audience and maximize ROI on advertising spend.

Publishers need to be able to gain advertisers’ trust through transparency of data and at the same time have interest in monitoring advertising’s performance on their properties to identify the best ad formats and implementations as well as the best fitting ad inventory.

The problem is that often advertisers and publishers don’t use the same measuring tools and systems. However, there are key metrics, common units of measurement that allow both parts to talk the same language.

The fundamental questions that digital marketing measurement tries to answer are essentially two:

Audience: who are they?

Reach: how many are they?

Let’s start with the second.


There are several tools to measure reach, such as Google Analytics and Website logs, but let’s firstly have an overview of the most common key metrics in use.

– Hits

Hits have been ironically redefined by many as “How Idiots Track Success”. Let’s see why.

A hit is a request to a web server for a file. Each file sent to a browser by a web server is an individual hit. When a user visits a web page with their browser, not only will the browser send a request to the server hosting that page for the html file itself, but it will generate also an individual hit (an individual request) for each file that page is made of (images, embedded videos etc).

Because of this, hits can’t be considered reliable to measure web traffic, because for a single page view (or page impression), many hits may be registered.

– Page views or Page impressions

A page view, instead, is a request to load a single web page as a whole, regardless of how many files the page is made of and how many hits are generated.

As explained in Google Analytics support website, it is important to note that “if a visitor clicks reload after reaching the page, this is counted as an additional pageview. If a user navigates to a different page and then returns to the original page, a second pageview is recorded as well.”

– Visitors or unique visitors

Visitors or unique visitors are the number of unique individuals visiting a site (as a whole, not specifically a single web page within the site).

A unique visitor is usually identified via an anonymous cookie. It is important to note that this is usually only a close estimate, not an exact measurement, as there are many scenarios that can cause imprecision, such as:

– Two people visiting a site using the same computer and same browser will be identified by the same anonymous cookie and therefore counted as one unique visitor.

– One individual visiting the same site from two different computers will be issued a cookie on each computer and therefore counted as two unique visitors

– An individual visiting a site from the same computer but using two different browsers will be counted as two different visitors, because each browser will be issued its own cookie.

– If an individual visits a site and is issued a cookie, but then visits the same site again after having cleared their cookies, that same individual will be issued a new cookie and counted as a different visitor on that second visit (therefore counted as two different unique visitors).

Unique visitors are calculated over a defined period of time or reporting timeframe (e.g. day, week, month, years etc), but it is essential to note that individual visitor counts from different periods can’t be simply added up to get an overall visitor count.

If I count my visitors on a daily basis, for instance, I can’t simply add up my visitor counts for the 7 days of the week to get the overall weekly visitor count. An individual tracked as unique visitor on Monday could return on Tuesday and be tracked again as unique visitor, therefore my weekly addition would result in an inflated number.

Visits or sessions

In web analytics terms, a visit (or a session) starts when a visitor starts interacting with a site (typically the first page view by the visitor) and ends when a specified period of time elapses since the last interaction of that visitor with the site.

In most web analytics tools a visit ends after 30 minutes of visitor’s inactivity on the site, but this can vary and can be personalized.

If the time limit of a single visit/session is set at 30 minutes and one visitor leaves your site but then returns to it within 30 minutes, that will be counted as one single visit/session, whereas if the same visitor leaves and return after 30 minutes, that will be counted as a second visit/session.

Differently from unique visitors, the number of visits to a site can be summed up across different periods of time to get an overall visit count.

It is important to underline that a visitor accessing a site for the first time is counted as a new visitor and a new visit. If the same visitor returns to the site a second time (without having deleted their cookies and accessing the site from the same computer/browser/device) after the set session time has elapsed, that will be counted as a new visit, but not as a new visitor.

New and returning visitors

We have mentioned how counting unique visitors can only be an approximation because of the problems related to cookie tracking. This is also a reason for caution in the assessment of new visitors vs returning visitors.

If I visit a website and later on return to it (in a different visit/session), the cookie that was issued to me on my first visit identifies me as a returning visitor.

However, if I do that after having deleted my cookies or if I visit that site from a different device/computer or a different browser, I will be issued a new cookie and identified as a new visitor. This is why statistics on new visitors tend to give overestimated numbers.

In conclusion, there is a good range of key metrics to measure web traffic, but these metrics don’t say much about who our audience is, who is behind the browser. This will be the topic of my next post.

Paid, owned and earned media: the POEM categorization

A popular way to categorize media in the marketing industry today – and perfectly valid for the digital world in particular – is denoted by the acronym POEM and distinguish between paid, owned and earned media types. Sean Corcoran (Forrester Research) has outlined the main features of each category highlighting the different roles and specific challenges of each type.


 In Corcoran’s summary, paid media are defined as those where the “brand pays to leverage a channel”, or as Dave Chaffey puts it, those “where there is investment to pay for visitors, reach or conversions […]”.

This is the model of traditional offline media such as television and print, for instance. Typical samples of paid media types in the online world are display ad networks, affiliate marketing and paid search.

Immediacy and scale are the main benefits specific to paid media, whereas the main challenges lie in the cluttered nature of the advertising landscape, the consequent declining response rate and the poor credibility of a type of media towards which consumers have grown more and more resistant.

Corcoran notes how “paid media is shifting away from the foundation and evolving into a catalyst that is needed at key periods to drive more engagement” and is meant to “feed” owned and “create” earned media.

For example, a time-based display advertising campaign run in a key period – e.g. the launch of a new product or a special offer – should eventually support the longer-term efforts of an online presence built up through the company website, as well as generating social buzz and earned interest.


Typical examples of owned media – a “channel a brand controls” – in the digital world are websites, blogs and mobile appsSocial media presence – such as on Facebook, Google+, LinkedIn or Twitter – is also seen as at least partially owned media.

According to Corcoran, the role of online media is to build “longer-term relationships with existing potential customers and earn media”.

Owned media are a different type of investment, yet an investment all the same, but it’s important to underline how they are meant for a longer-term perspective, as they lack the immediacy of a targeted paid media campaign.

The versatility of owned media enables brands to tailor content for both broad and niche audiences; even though not immediately scalable, this type of media can provide – in Lee Odden’s words –  “long term growth benefits without corresponding growth in costs”.

From a brand’s perspective one of the most appealing aspects of owned media is the fact it allows direct communication with their audience in a controlled environment. However, the key element is not so much in that control per se, but in the opportunity to provide the audience with high quality content. The challenge for marketers, as described by David Germano, lies in shifting “from content that represents their messages to consumers to content that audiences themselves are looking for”.


Traditionally the term “earned media” refers to PR publicity generating brand awareness and media presence without a direct investment in paid advertising.

In the online world, it is used in particular to describe word-of-mouth through social media as well as viral marketing. As much as it is impossible to really predict what will go viral and what won’t, brands can proactively stimulate social buzz by engaging with online communities and customers through social media.

As Corcoran puts it, it’s about letting customers “become the channel”.

Corcoran lists credibility as one of the main benefits of earned media, as opposed in particular to paid media. Customers are more likely to take the word of other customers rather than the word of the brand itself.

On the other hand, one of the most disputed aspects related to brand engagement with social media is the risk of opening the gates to spurned criticism and bad PR. However, many have identified the increasing need for brands to be more “human” and able to be open and honest about their flaws.

The recent episode of Bodyform’s ironic (and slightly mean) comeback to a facebook post about the deceptiveness of tampon ads comes to mind.

Some (e.g. Odden) classify social media as a different category altogether, defined often as “shared media”, since it hinges on the interaction between brand and consumers.

Like owned media, earned media are also an investment even though there is no direct money exchange. As Gavin Llewellyn notes, the investment is one of “time, effort and resources”. Llewellyn also adds that “earned media isn’t easy to achieve and for many paying for an advertisement is quick, simple and often easier to calculate ROI.”

In conclusion, the different media types can support and sustain each other, therefore brands can benefit greatly from combining these different media in an integrated strategy that takes into account their different specificities.

Why measure?

In my last post I mentioned the famous Lord Leverhulme quote – “half the money I spend on advertising is wasted, the problem is I don’t know which half” – to refer to the fact that digital technology has provided marketers with more powerful ways to measure the effectiveness of their campaigns.

The importance of measuring systems may be obvious and self-explanatory, but it can be useful to still ask yourself why they are beneficial.

In very broad strokes, measurement is a necessary condition to quantify results and enhance future performances.

In addition, measuring allows you to:

– Get to know your audience, see how visitors interact with your brand, channels and properties
Distinguish between more and less successful marketing initiatives
Quantify ROI on marketing spend

Measuring a digital marketing campaign depends on the budget, the tools/technology and the resources available; however, regardless of these variables, a key question is what is to be measured.

There is no one-size-fits-all approach. It’s up to you to decide what to measure and what to ignore and what KPIs (Key Performance Indicators) to consider. It’s important to know what you intend to do with that data, what decisions it would facilitate. In Avinash Kaushik’s words, before the data and the tools, what is essential is “structured thinking about what the real purpose of the campaign is”.

With that respect, it is crucial to be familiar with the specificities of the different media which we avail of in today’s diversified digital marketing landscape.

These different media will be the topic of my next post.

The digital marketing landscape, an overview.

Before approaching the core topic of this blog – digital marketing measurement – I would like to give an overview of what the digital marketing landscape looks like and define what digital marketing is in relation to traditional marketing.

First of all though, what is marketing?

Many definitions have been given; the following – by Kotler and Armstrong – is often referred to:

“[Marketing is] the process by which companies create value for customers and build strong customer relationships in order to capture value from customers in return.

In other words, marketing is essentially about trading value. Marketers provide consumers with something of value so that the consumers trade something they value in return, be it money, time, personal information etc.

Now, is digital marketing something different?

I would say no. The same definition stands true in the digital world. Digital is just a different channel, but it doesn’t alter the core of marketing dynamics, which are still about the same value exchange.

This doesn’t mean marketers should keep doing online what they have been doing offline without reflecting on the specificities of this new channel.

What has the advent of digital changed in the marketing landscape, then?

Many things could be mentioned, but I’d like to point out two main aspects:

More fragmented landscape

The fragmentation of the marketing landscape hasn’t started with digital, but digital has definitely increased it exponentially. People nowadays are likely to be texting on their mobiles while they walk past a billboard; they are likely to be online on their laptop or tablet while they’re watching TV. In other words, modern technology has made us more and more multitasking while simultaneously giving marketers more and more channels to reach us, but we’re easily distracted. Winning our full attention is an altogether tougher challenge.

More power to the consumer

Traditional Mass-media communication is one-to-many; digital communication is more and more based on the many-to-many model of social media.

Traditionally brands would have full control on the flow of information and the messaging of content; nowadays consumers have more control on what information they are receiving and have more means to actively seek out content. What’s more, consumers are media producers too; the power of communication is shared.

With an increase in consumer control and engagement, marketing communication has changed from a model of “interruption” to a model of “participation”; marketing is less about “saying things to people” and more about “doing things for people”.

In a more fragmented and “democratic” communication system, not only is it important for marketers to pitch the right channel for their message, but it’s also more and more important to be able to organize their message through the various channels, as the growing success of transmedia storytelling demonstrates.

If on the one hand digital technology has made the marketing landscape more complicated, on the other it has empowered marketers with new tools and means to measure the effectiveness of their campaigns more accurately.

Lord Leverhulme is known to have said: “half the money I spend on advertising is wasted, the problem is I don’t know which half”; perhaps he would have appreciated the digital marketing measurement tools available today, but this is a story for another post.

Post Navigation