Media Measurement and main models.pdf

CONFIDENTIAL & RESTRICTED OFFERING CONTENT EXPERIENCE When a measure becomes a target, it ceases to be a good measure COPYRIGHT © 2019 ALL RIGHTS RESERVED. CONFIDENTIAL. Media Measurement and main models CONFIDENTIAL & RESTRICTED PAID MEDIA CAMPAIGN STRUCTURE Cross Strategic Foundamentals & Trends K...

CONFIDENTIAL & RESTRICTED OFFERING CONTENT EXPERIENCE When a measure becomes a target, it ceases to be a good measure COPYRIGHT © 2019 ALL RIGHTS RESERVED. CONFIDENTIAL. Media Measurement and main models CONFIDENTIAL & RESTRICTED PAID MEDIA CAMPAIGN STRUCTURE Cross Strategic Foundamentals & Trends KPIs & Attribution Models Ecosystem Media Planning & Buying Innovation Vertical / Deep-dive Search Display Video Paid Social Media CONFIDENTIAL & RESTRICTED GUT FEEL MARKETING IS A BAD PLAN FOR 2024 Conventional methods are incomplete, misleading, slow and generally suboptimal Data science + Advanced Analytics = Competitive advantage CONFIDENTIAL & RESTRICTED FIRST: KPIs RECAP!KPIs definitions IMPRESSIONS CONFIDENTIAL & RESTRICTED Impressions represent the number of times an ad appears on a web page. An impression is simply a record of views. REACH Measuring the success of your display advertising depends on the types of campaigns you are running and your chosen KPIs. This metric demonstrates how many people see your ad. It shows the number of unique views a particular ad receives. ENGAGEMENT RATE Usually displayed as a percentage, the engagement rate shows how many people interact with an ad. This could be something as simple hovering over an ad, or more complex, like filling in a search-banner. CTR - Click-through rate The total number of people who click on the ad after seeing it. Cost-per-click (CPC) Total cost divided by the number of users clicking on the ad. Return on investment (ROI): The difference between the total revenue generated from your campaigns and total cost of running it in order to calculate the net profit. CONFIDENTIAL & RESTRICTED The user becomes increasingly complex, increasing the number of points of contact before making the purchase. Before users makes a purchase of a product or service, they have gone through different points of contact, viewed ads on social networks, compared prices on Google and received an email, as well as the offline channels. At first glance it seems simple, but it rarely happens that a customer goes directly to a website to make the purchase. In a utopian aspect, we could consider the buyer journey flat and linear, from beginning to end, but this is a biased view of reality. CONFIDENTIAL & RESTRICTED Which marketing or media channels have had the greatest impact during the purchase decision process? Multiple channels and messages have been involved in this process, these being responsible for the purchase decision. Attribution marketing uses attribution models to assign a value to the different touch points that the consumer has passed through. CONFIDENTIAL & RESTRICTED What is an attribution model? An attribution model is a set of rules by which a certain value is assigned to the different channels that a user has gone through before performing an action that we interpret as a conversion. CONFIDENTIAL & RESTRICTED There are different models: It is very important the attribution model you choose, because there are many, and everything depend on it, even using the same baseline data, the results will be completely different. CONFIDENTIAL & RESTRICTED How did my channels perform? Which were my most profitable channels? Which creatives drove engagement? What channels delivered results earlier in the funnel and which ones delivered results later? Which creatives worked the hardest? …. CONFIDENTIAL & RESTRICTED CONFIDENTIAL & RESTRICTED How you split up the credit? Some channels are openers, closers, some other neutral, others assist…. Benefits of attribution analysis Look for the effective investment (budget optimization): What is really working and what is not. This is the main goal of using an attribution model. Channel Synergy: This information allows us to optimize the channels and determine the importance of each of them, the keywords that attract users or which channel precedes and proceeds the maximum experience. Development and creativity: It gives the opportunity to evaluate the most creative elements of the campaigns and the most attractive ones with the aim of perfecting the message, the visual elements and knowing how and when to communicate with the users. CONFIDENTIAL & RESTRICTED CONFIDENTIAL & RESTRICTED MAIN ATTRIBUTION MODELS CONFIDENTIAL & RESTRICTED MAIN ATTRIBUTION MODELS We can establish two main types of attribution models: Single touchpoint. Multi-source or multi-touchpoint. CONSIDER THAT: The initial attribution models did not consider all touchpoints as critical to conversion. The most advanced models understand the interactions between different marketing channels and between devices as the key to a faithful attribution system. There are also external factors like weather, better price/deal offered by the competition etc. that are not taken into account by attribution models. CONFIDENTIAL & RESTRICTED 1. Single touchpoint attribution 1/3 These models assign all credit to a single touchpoint, being the first or last or the Lead-Conversion. 1. Last touchpoint or last click. This type of attribution grants full attribution credit to the last touchpoint that the consumer interacted with before converting. It does not take into account the past interactions you have previously been through, thus overlooking key information about other marketing channels. This model is the most popular and the one that is predetermined in the Google Analytics account. If conversion is the main objective of our campaigns, then the last touchpoint model is the easiest to use but of course, it ignores all the steps that were taken up to the point of conversion. Use it for understanding channels that directly impact conversions CONFIDENTIAL & RESTRICTED 1. Single touchpoint attribution 2/3 These models assign all credit to a single touchpoint, being the first or last or the Lead-Conversion. 2.First touchpoint or first click. The "First Touch" attribution model assumes that the consumer converted after the first ad they viewed. Grants full attribution at the first touchpoint, regardless of what happened next, so 100% of the success is attributed to it. We could say that it is the perfect and simple method, easy to configure and focuses on brand awareness. Nevertheless, it is also susceptible to channel bias, technological limitations and user behavior, since customers have gone through different points before converting and contact with them nurtures the purchase decision. Use it to understand which channels impact brand awareness CONFIDENTIAL & RESTRICTED 1. Single touchpoint attribution 3/3 These models assign all credit to a single touchpoint, being the first or last or the Lead-Conversion. 3. Lead-Conversion click The other common single-touch model is the “lead conversion click” attribution model. This is where 100% of the revenue credit is given to the action that drives the visitor to become a lead (i.e. give you their contact information by filling out a form). For example, if an anonymous visitor comes to our website multiple times (browsing your site) and downloads content on their fifth visit, the channel that drove the fifth visit would receive 100% of the revenue credit when that lead becomes a customer. All of the other visits would be ignored and would receive 0% of the revenue credit. The tricky thing here is that both the First Click attribution model and the Lead-Conversion Click attribution model are sometimes called “First Touch.” When it comes to many marketing automation systems, their view of first touch is typically when the lead was created, not their actual first visit to the website. But when it comes to web analytics that are cookie-based, First Touch attribution is viewed as the same as First Click attribution. Use it if your campaign goal is to increase leads. CONFIDENTIAL & RESTRICTED 2. Multi-source or multi-touch attribution 1/6 This model attributes credit to each channel for being part of the user's final conversion. It represents the entire buyer journey and takes into account multiple touch points that include everything. 1.- Linear Model This model weighs equally each of the touchpoints through which the consumer has passed. Whether the first contact, the last one or any event in between, everything is treated the same. The linear attribution model is a step forward compared to single touchpoint attribution, giving us more insight into what has happened. The intermediate stages can be as important as the first and last, all of them contribute to the purchase in an equitable way. However, giving each impact the same percentage of credit is almost unrealistic when it comes to conversion, as we know that there are points that affect conversions more than others. Use it to see all the channels and campaigns that impact conversions CONFIDENTIAL & RESTRICTED 2. Multi-source or multi-touch attribution 2/6 This model attributes credit to each channel for being part of the user's final conversion. It represents the entire buyer journey and takes into account multiple touch points that include everything. 2.- Time Decay The model decreases the amount of credit given by half after a set amount of time, with more credit given to the most recent touchpoints. Its objective is to be able to recognize that the different points have a different value and it seems logical, considering that the last points of contact are represented in a funnel in the middle and lower part. Although this model leaves aside that the first touches can be very influential. Imagine that you register for a service asking for a demo, and you buy six days later and the day before the purchase you had clicked on an email with a blog post. Should that blog really get more credit than the demo? Probably not. Use it if you have a long sales cycle CONFIDENTIAL & RESTRICTED 2. Multi-source or multi-touch attribution 3/6 3.- U-shaped The U-shaped or position-based model emphasizes two key contact points: the first and the last contact by assigning 40% of the credits to each of them. The remaining 20% is divided between the contacts that occurred in an intermediate way. It makes sense that we use this model to measure the effectiveness of campaigns that start with PPC (pay per click) ads and end with product and pricing page views. Use it if you place higher value on awareness and conversion CONFIDENTIAL & RESTRICTED 2. Multi-source or multi-touch attribution 4/6 4.- W-shaped It is similar to the previous model but includes an additional touchpoint: the creation of opportunities. The highest value is attributed to the three main stages of the buyer journey: visit, lead creation and conversion. Each of these receives 30% credit and 10% is divided among the others. Use it if you're focused on lead generation. CONFIDENTIAL & RESTRICTED 2. Multi-source or multi-touch attribution 5/6 5.- Z-Shaped or Full Path Attribution. Instead of just three key points as the W model provides us, it adds one more stage. Each of the touchpoints receives 22.5% credit and 10% is divided equally between the remaining points. One of the benefits of the full path attribution model is that it gives the same weight to initial marketing activities as it does to subsequent follow-ups of the sale opportunity. Use it to understand which channels impact brand awareness CONFIDENTIAL & RESTRICTED 2. Multi-source or multi-touch attribution 6/6 6.- Data Driven, Algorithmic or personalized attribution It refers to those models where a percentage of the sale is assigned to each interaction based on the value it brings together with the rest of the interactions. These models are based on the study of historical data and uses algorithms (Markov model, Shapley, Bayesian model , Regressions, Game Theory….). It is one of the most advanced ways to collect and deduct attribution data. By analyzing data from existing customers, you can determine which of the channels had the greatest impact. A company can have a personalized and unique model for its business. This is a difficult and time-consuming model to build, but it undoubtedly provides the most accurate representation of the impact of marketing actions on the customer journey and final conversion. Use custom models if you have long and very complex customer journeys CONFIDENTIAL & RESTRICTED DIGITAL ATTRIBUTION: Focus types example Attribution model User path example: Social Media Display SEM Email Conversión Which channel is generating the conversion? Last Click Email First Click Social Media Lineal 0,25 of the conversion to each channel Time Decay Greater weight to the channels closest to the conversion(0,40; 0,30; 0,20; 0,10) Data Driven All channels take a part of the conversion depending on the algorithm CONFIDENTIAL & RESTRICTED GOOGLE ANALYTICS ATTRIBUTION MODELS IN GA There are three types of attribution models available in the Attribution reports in Google Analytics 4 properties: - Cross-channel rules-based models: Cross-channel last click, first click, linear, position-based, time-decay - Ads-preferred rules-based model: attributes 100% of the conversion value to the last Google Ads channel that the customer clicked through before converting. - Data-driven attribution: Data-driven attribution distributes credit for the conversion based on data for each conversion event. It's different from the other models because it uses the account's data to calculate the actual contribution of each click interaction. Each Data-driven model is specific to each advertiser and each conversion event. To find the Attribution reports, click Advertising on the left. Under Attribution, click either Model comparison or Conversion paths. All attribution models exclude direct visits from receiving attribution credit, unless the path to conversion consists entirely of direct visit(s). CONFIDENTIAL & RESTRICTED CONFIDENTIAL & RESTRICTED CONFIDENTIAL & RESTRICTED CONFIDENTIAL & RESTRICTED CONFIDENTIAL & RESTRICTED Planning a project: Data quality Before starting a digital attribution project (or when we are using the results) we have to reflect on some relevant aspects: Data on digital media - Is all activity being measured? Not only click, also impressions - Is visibility being measured? (When possible) If we know that an impression has not been seen, it should not be part of the model… - Does the trafficker collects all necessary data for analysis and decision making? (Naming Convention of an AdServer or the UTM of analytical tools etc…) Data on the digital assets - what happens on the site is correctly measured? (i.g.: artificial visits should be excluded) - Do we have all the conversions? What about non-conversions? (Alternative channels) - Is the implementation agile enough? Do we have a stable environment? How to choose the right attribution model We have to take into account different variables to choose the right attribution model, since very different campaign and action perspectives can be obtained depending on the model. 1. Consider the time it takes for a user to convert 2. Single Touchpoint models do not take into account the entire purchase cycle, as the longer the conversion takes, the more complex it will be, including multiple visits or even comparison with competitors. 3. Take into account the number of active channels in your sales cycle, including the offline aspects. The fewer channels that participate, the easier the model you can use will be. 4. Long purchasing processes require algorithmic or W-shaped models. 5. Establish what you are trying to measure: it is precisely these objectives (conversion or attracting new customers…etc.) that will guide you to define the correct model. 6. The ideal situation is to identify the user even on different devices: cross device tracking and attribution 7. Try: the more you try, the closer you will be to the correct attribution model. CONFIDENTIAL & RESTRICTED AMADEUS DASHBOARD CONFIDENTIAL & RESTRICTED Looking toward the future: #trends Measuring cross-everything Marketers must be able to measure cross-channel, crossplatform, cross-device, and cross-media. This will give marketers an even more complete and accurate picture of how people really shop and buy. Proving business impact Marketers will increasingly focus on metrics that show a direct impact on business outcomes. As technology continues to advance, marketers will gain access to new measurement solutions that help them see the impact of their marketing on outcomes like revenue and profit. CONFIDENTIAL & RESTRICTED Relying on AI Marketers will rely more and more on machine learning and automation to help them manage the explosion of data signals available to measure. Machine learning will be required to help rationalize all of the data available and identify the specific, actionable insights they can apply to their marketing to drive growth. Which Attribution Tools are best for Attribution? There are many tools and vendors in the market that can provide powerful attribution platforms but the approach to the tools and the strategy behind is more important than the tools themselves. Also, this space is still really young. All vendors admit that the way date ranges in the model work, relating the time period of ad spend vs. conversion completion, leaves a lot of room for error. It is important to emphasize that all of the tools available will be best used for the optimization of paid channels. For some of our larger clients, attribution vendors would not provide what they’re looking for, because they want something more custom. The alternative is to build your own model on top of a data warehouse and a customer data platform. Some of them are: CONFIDENTIAL & RESTRICTED CONFIDENTIAL & RESTRICTED Additional readings IF YOU’RE USING A SINGLE TOOL FOR MEASUREMENT, THINK AGAIN by Think with Google https://www.thinkwithgoogle.com/marketing-strategies/data-and-measurement/mediameasurement-tools/ WHAT’S NEXT IN MEDIA MEASUREMENT? by Nielsen International Media https://www.nielsen.com/ssa/en/insights/article/2019/whats-next-in-media-measurement/ OVERVIEW OF ATTRIBUTION MODELING IN MCF Assign credit for sales and conversions to touchpoints in conversion paths by Google Analytics https://support.google.com/analytics/answer/1662518?hl=en ABOUT CONVERSION WINDOWS by Google Ads Help https://support.google.com/google-ads/answer/3123169?hl=en GOOGLE ANALYTICS ATTRIBUTION MODEL EXPLAINED https://measureschool.com/google-analytics-attribution/ CONFIDENTIAL & RESTRICTED Exercise CONFIDENTIAL & RESTRICTED 1. 2. 3. 4. 5. Which 3 metrics do you consider the most useful in the table? Why? Why do you think Instagram has the worst conversion rate? Which source/medium has the worst performance? Which media channels would you use to acquire new customers? What do you think is the source with the greatest opportunity for optimization? Consider that it is an ecommerce website. Make your assumptions and answer to the 5 questions considering the table below. Think in the whole user journey and cross-device.