Data-Driven Attribution Modelling Explained (Part 2)

Spread the love


One-click purchases are a rarity, rather than a reality for most businesses, and yet the standard attribution models imply that this is how purchases happen online. But companies, who are crediting only the last touchpoint in analytics for overall marketing success, are missing out on 20%-40% of potential ROI. Knowing more and guessing less about your customers’ journeys can help you recuperate that money and even multiply your marketing gains. That’s what data-driven attribution modelling is designed to accomplish.

What is data-driven attribution?

As mentioned in our previous post, attribution models that follow pre-defined rules (e.g. assign credit to first/last click) do not provide full visibility into your marketing campaigns. They are decent “patch” solutions to track certain activities, but they fall short when you want to dig into the complex customer journeys of today.

Google offers a Data-Driven Attribution (DDA) model which accounts for the importance of every touchpoint a prospect goes through before converting and determines which marketing actions played a role in the process. However, to start using this model, your business must first meet the general eligibility requirements. To meet these, you must have:

  • A Google 360 Analytics account which costs £115,000 per year
  • Either E-commerce Tracking or Goals set up
  • At least 15,000 clicks on Google Search and at least 600 tracked conversions within 30 days.

Unfortunately, this isn’t an option for many businesses, but there is another way.

Develop your own data-driven attribution model

Yes, you can develop your own model instead, based on GA data. We have built these for several of our customers and they are highly effective. And, unlike the Google model, with ours, you only need a minimum of 100 sessions per day, including traffic from all channels, which makes it much more accessible for many companies.

These models comb through different conversion paths and identify the number of touchpoints in different sequences, the order of exposure, creative assets used and several other factors to give you a complete and actionable view. The model uses the conversion path data from MultiChannel Funnels, as well as path data from customers who don’t convert.

We can then set up predictive analytics modelling that will supply your business with insights and valid predictions in near real-time. This way, you can identify winning campaigns and sequences, and make better decisions on-the-fly.

What are the benefits?

Chart showing how many conversions were attributed by channel and distribution model

Custom data-driven attribution enables you to create a personalised model that will depict the actual state of affairs for your business. You can add, track and optimise as many channels and touchpoints as you need to fully document the different types of customer journeys.

READ ALSO  The Definitive Guide to Data-Driven Marketing

The best part is that all insights are connected, which provides a detailed look both within and across channels. Specifically, you can unlock the following benefits:

  • Improved budgeting: learn how each channel contributes to your set goals and optimise your spending for different campaigns with granular precision.
  • Cross-channel impact analysis: lift and funnel stage reports will reveal how different channels and marketing actions affect one another.
  • Unified view: set up a custom set of metrics for all channels, aligned to your business goals.
  • Reporting flexibility: generated reports can be viewed in multiple ways, allowing different teams to instantly access the insights they need.
  • Performance over time: estimate and predict customers lifetime value, drop off chances and multiple other factors that impact your business’s bottom line.

Ultimately, data-driven attribution allows you to give proper credit to previously hidden actions, such as conversions that came from non-branded keywords or from mobile devices (after a visit from a desktop), as well as exercise more precise control over individual campaigns/channels. Additionally, you gain access to predictive analytics insights, showing you the scope of change you will achieve when trying strategy A, B or C.

Several brands in different industries are already seeing great results after switching to DDA:

  • Select Home Warranty in the US witnessed a 36% increase in leads and a 20% drop in cost-per-conversion.
  • Medpex in Germany generated 29% more conversions while reducing cost-per-conversion by 28% using data-driven bidding on Google Adwords.
  • H.I.S., a global travel brand, combined DDA with Smart Bidding and Dynamic Search Ads to boost conversions by 62% without increasing the cost-per-conversion.

What do you need to implement a data-driven attribution model?

1. Google Analytics set up and configured

As mentioned already, you don’t need to be a Google Analytics 360 customer if you want to benefit from data attribution modelling with our help. But there are still a few technical requirements you need to meet:

  • Active Google Analytics and Google Tag Manager accounts
  • Either E-commerce Tracking or Goals set up
  • Receive at least 100 sessions per day, including traffic from all channels.

Implementing data-driven attribution at a lower daily traffic threshold is still possible, but you won’t be getting complete value out of it at that point.

2. High data maturity

Your analytics will only be as good as your data is. Businesses that are able to gather large, consistent and high-quality data sets across the channel mix will benefit most from data-driven attribution. Specifically, you’ll want to ensure that your data is:

READ ALSO  What You Need To Know About Starting a Web Design Business – Part 1

Types of data: focused data, systematic data, low data and chaotic data

Source: Google Attribution 360 White Paper

  • Focused: it’s collected from different channels and is relatively easy-to-access.
  • Systematic: you are able to collect relevant data from an online/offline marketing mix, and have respective processes in place for that.

Remember: all the big data stored in your systems will have to be operationalised and prepared for further analysis. In fact, that’s what we tend to focus on during the first two months of working with a customer.

3. Aligned KPIs and business goals

Data-driven attribution works best when your KPIs and goals are compatible across channels and departments. For example, if your PPC campaign is geared towards increasing the webinar leads, your Facebook ad campaign goal should be the same.

But there are different ways to measure those goals, right? What’s great about data-driven attribution is that it allows you to look beyond the vanity metrics such as clicks or shares on social media and focus on post click insights (e.g. visits) instead.

Not every social media campaign you run may be aligned with a specific outcome, such as sales. An Instagram campaign you run might be tailored to create brand awareness in a new market. But measuring that “buzz” can be problematic. With a DDA model, however, you can effectively capture social media traffic that never generated a conversion, and analyze how it had helped the performance of other channels.

Should you implement DDA?

Depending on the company size and how much value you see in data, a data-driven attribution model can be the tool you need to help you answer important marketing and business questions, such as:

  1. Which channels contribute more towards conversions?
  2. Do any campaigns not provide ROI?
  3. Are there particular affiliates that increase the probability of conversion?
  4. Which referral sites are crucial to the user journey?
  5. What will happen if the PPC budget gets reduced?



Source link


Spread the love

Leave a Comment

Your email address will not be published. Required fields are marked *