What are the most effective cross-device attribution modeling practices for digital marketers?

(!) When defining your goals and KPIs for cross-device campaigns, it's crucial to consider the unique user behavior associated with each device. For instance, if tablet users tend to have higher on-page interaction but lower conversions, your KPIs for tablet campaigns might prioritize engagement over direct sales. Conversely, if desktop users often convert after viewing a product on mobile, then your attribution model should give weight to both touchpoints.

If you would like to improve your Key Performance Indicators (KPI), then do the following first: Keep People Informed Keep People Interested Keep People Involved Keep People Inspired

Defining your goals and KPIs is a huge first step. Everyone’s top KPI is going to be Opps closed or products sold. However, as marketers, we have to go a little higher up the funnel than that, and attribution is a key part of that. Last Click will show you the touch points that pushed someone over the edge; that’s helpful to know, but it’s only a few tiles in the mosaic. I prefer time decay as it will show you almost every touch that an individual went through before purchasing. Time decay allows you make informed optimizations to your top of funnel campaigns. Last click will only focus on the very bottom of the funnel. So if you’re only running BOFU campaigns, last click is your bread and butter. Otherwise, try out Time Decay.

Focus on Business Goals: Always tie your attribution efforts to your business objectives. What matters most is how these insights can be used to improve marketing strategies and ROI.

You can't decide on actions before your goals. Firstly understand your clients/users journey and align KPIs for each step on the way depending on the funnel size. Then separate them into different column based on funnel stage. Now it is the time to understand What to measure and how. For instance, If I am having a app-install journey but my first touch is awareness and I choose Shorts. Then I will put as KPIs engagement and time watched. This will be the explanation of the app in a sense of which problem it solves.

Understand the different attributions first. When picking an attribution model is like choosing your marketing strategy. Google through every attribution model that exists and understand the particularities. I usually advise beginners to start with data-driven until they understand their users.

Use a unified customer view, a single view of each customer that includes data from their interactions with your brand across different devices and channels. This data helps create more personalized marketing campaigns. Machine learning can help analyze cross-device attribution data to find patterns and insights that are hard to find manually. This info can help to optimize marketing campaigns and improve ROI. The best model for you will depend on your goals and objectives. Use a variety of attribution models to get a more complete picture of your marketing performance. Collaborating with other departments like sales and customer service is vital to understanding the customer journey and the impact of marketing efforts on the business.

Effective cross-device attribution modeling in digital marketing involves: Model Familiarity: Understand various attribution models and their pros and cons. Goal Alignment: Choose a model in line with campaign goals and KPIs. Data Suitability: Consider data availability when selecting a model. Cross-Device Focus: Opt for models like linear, position-based, or data-driven for cross-device campaigns. Budget Optimization: Use the chosen model to allocate budgets based on device and channel contributions for better ROI.

Given the increasing reliance on machine learning and artificial intelligence, the ability to choose an attribution model won't be available for longer. It all will come down to data-driven and statistical modelling-based methods. Why? Because it gives more freedom to ad platforms and is more reliable. First click, last click, linear, position-based, and time-decay, all of these are biased to specific touchpoints. Data-driven modelling appears more logical here.

To identify the right touchpoint, you'll need to go back to the basic measurement model, demographic data. It is important to know where these interactions come from and if there is a pattern such as sex, age or residence that intervenes when they call to action. Then, you can consider using multi-touchpoint attribution models, depending on your audience and their conditions (those that you already identified). From that stage, each model gives different weight to various touchpoints in the user journey, allowing you to better allocate credit for conversions.

Cross-device tracking is easier said than done. With the recent changes in privacy norms and how third-party tracking works, it has become even more serious. Use whatever tools or system you can to have cross-device tracking.

That is the most difficult one. Nowadays with all the GDPR layering we can't really control these. Use the tools but also use some speculation methods to understand hidden points.

When monitoring you need to have the specific KPIs for each journey as well as the over all KPIs of the whole campaign. When you start analysing your results you need to pay close attention to your assisted conversions. We often shut down a campaign that gives us the kick start, on in other words initiates the journey but it doesn't convert the end result. When this happens we see a massive reduction in results as the user journey is not really kick starting the way it used to.

In my experience, analyzing the results are very important to rethink on our strategy we developed at the initial point. The results may not be satisfying sometimes but it helps us to find what went wrong in the process.

Analyzing cross platform strategies can be tricky, and the attribution model you choose may not sometimes do justice to the data you see. One way to solve this is to classify the campaigns based on funnel-stage. For example, we've realized that Facebook's AI chooses the audience based on your goal, so if you're choosing your goal as "get more traffic" it will show the ads to people who are more likely to click (but click doesn't necessarily mean conversions.) And so, even for Top-Of-Funnel (ToFu) campaigns we choose conversion goal but look at the reach/traffic we have gained even though there aren't many conversions. We even add the word "ToFu" in the campaign name to identify it easily.

One of the biggest mistakes I see a lot of brands making is not tracking the correct data. Without a deep understanding of what works and what doesn't work, you are basically shooting in the dark hoping to hit something. Accurate data and a simplified understanding make it easier and quicker to start, end and scale campaigns.

Effective cross-device attribution modeling requires a holistic approach. Utilize multi-touch attribution models like linear or time decay, considering touchpoints across devices. Implement cross-device tracking tools to follow user journeys seamlessly. Leverage first-party data for accurate insights and invest in machine learning algorithms for predictive modeling. Regularly analyze customer behavior to adapt strategies. Prioritize user privacy and compliance while ensuring a seamless, integrated customer experience across devices.

It is a good idea to imagine and outline your user experience along with another 10 people. Find 10 different people in your company or friends and do an experiment. Understanding reactions on the journey is a crucial step. Usually I do this in a form of surveys and interviews. If your budget allows it, you can do simulations.

First Party Data - A very important key to this puzzle which is missing in this piece. As platforms are more and more moving towards privacy sensitive world, 3P attribution tools will become less and less effective. Gathering FPD on your assets (with user consent) has to be developed in the anticipation of these massive changes. Managing data from various sources and product surfaces will then ensure effective measurement. This data should be effectively linked to the user experience for a win-win strategy for users as well as the brand.

Effective cross-device attribution modeling for digital marketers involves leveraging data from various devices, such as TV, mobile, and more, to gain a holistic view of the customer journey. To achieve this, marketers should employ a multi-touch attribution approach that considers touchpoints across devices, using tools like probabilistic matching, deterministic data, and advanced machine learning algorithms. It's essential to prioritize data accuracy, privacy compliance, and a customer-centric perspective, enabling marketers to allocate budgets and optimize campaigns effectively while delivering a seamless, personalized user experience across devices.

??In today's fragmented digital landscape, understanding consumers' journey across multiple devices is paramount. Best Practices: ?Unified ID Solutions: Employ systems that track users across devices seamlessly. ?Probabilistic Modeling: Leverage statistical techniques to predict user behavior patterns. ?Look at Data Holistically: Integrate data from all touchpoints for a 360-view. ?Prioritize Privacy: Ensure data collection is GDPR, CCPA, and other regulations compliant. ?Iterate and Optimize: Continually refine your approach based on insights. Embrace these practices to better understand your customers' journeys and enhance your marketing ROI.

Remember, no single attribution model fits all scenarios. It's essential to understand the business's objectives, available data, and specific audience behaviors to choose the most appropriate model. As digital marketers, we must also ensure that our tools and platforms are cloud-based and accessible from anywhere, allowing for real-time adjustments and optimizations.