10 Tips for Better Attribution Modeling

Multichannel marketing increases outreach but it also complicates the customer journey. Conversions rarely happen at the introductory touchpoint and understanding which channel led to the final sale is vital for investing resources properly. 

Giving credit where it’s due among various marketing channels is called attribution modeling. Read on to find out everything you need to know about it and how you can perfect your model to highlight the most valuable parts of your sales journey.

Marketing Across Channels

To reach modern customers, companies have to communicate on a seemingly endless number of channels. It’s not uncommon to have a social media campaign, digital advertisements, traditional ads, media plugs, and website content running simultaneously. 

As people increasingly access the internet from mobile devices like smartphones and wearable tech, marketing communications have to become more agile to find receptive audiences where they are. Understanding the target audience and where they’re most likely to engage with promotional material makes advertising efforts more effective and saves the company wasted money in failed attempts.

One big downside of juggling so many marketing channels is that it can be difficult to optimize advertising strategy as a whole. Don’t be surprised if more than one avenue is successful at drawing customers and getting conversions. When that’s the case, how can marketing professionals and businesses know where to keep directing their advertising budget with any confidence?

That’s why attribution modeling is so important. With the right metrics, businesses can see exactly where people are engaging with their marketing materials and whether that interaction leads to a conversion at that touchpoint. There are many attribution models, each with its own advantages in certain contexts.

Touch Points In Attribution Models

Generally speaking, there are two main categories of attribution models: single-touch or multi-touch. You could also separate them into direct and non-direct touches.

Just what is a touch when it comes to marketing? A touch, or touch point, is an interaction between a brand and a customer or potential customer. Common touch points include digital advertisements, branded content, social media posts, product reviews, and subscription renewals among many others. 

Certain touch points are designed to reach existing customers the way a subscription renewal or cart reminder sent via email usually does. Others are meant to entice potential users and entice them toward a COA like a subscription button or purchase. 

Regardless of which cohort they’re designed to attract, these touch points come together to form the customer journey, which describes all the interactions beginning from first contact with a brand to a purchase and extends to future return purchases as well. 

Companies organize their attribution models according to the stages of their most common customer journey(s) in order to dole out rewards and target their marketing budgets where such money will be most effective.

Types of Attribution Models

Many more specific kinds of attribution models lie within the broader single-touch or multi-touch categories. Here are some of the most prevalent models:

• First-Touch Attribution

This model gives full credit for conversions to the first touch. Pretty simple, right?

Although it might not be the best choice for a widely varied marketing strategy, first-touch attribution typically works great for campaigns that focus on just a few touch points anyway. For example, if you’re rolling out a social media campaign, you can get a pretty good idea of its success with a first-touch attribution when the social media profile is that first touch. 

• Last-touch Attribution

Counter that with a model that only considers the final touch point for conversion credit. While using last-touch attribution could unwisely ignore a large part of the overall sales journey, it also highlights a key section of it – that final conversion. 

• Last Non-Direct Click

Similarly to last-touch attribution, a last non-direct click model focuses on the conversion. But this model eliminates direct traffic, which is what occurs when people go directly to a site or app for purchases. If you want to measure how effective your marketing has been, it’s smart to eliminate those users who made purchases without going through any marketing channels.

• Linear Attribution

Rather than giving total credit to just one touch point, the linear attribution model splits credit among several. Businesses with sales journeys featuring many entry ways and customers who might engage with the brand several times before making a purchase tend to employ linear attribution models with great success.

• U-Shaped Attribution

A variation of linear attribution is U-shaped or position attribution. Credit is still split between every touch point but that credit is weighted to focus primarily on the first and last touch points. The logic behind this approach is that it shows businesses how people first get word, which touches keep their interest, and which ones tend to clinch that coveted final conversion. 

• Time Decay Attribution

If the focus is on long-term relationships, companies frequently use a time decay attribution model, which puts more weight on the touch points that are closer in time to the conversion. While that does leave out marketing methods near the top of the sales funnel, it also shows how the sales journey builds up to a conversion.

• Tailored Models

In some cases, even tweaking the models we already mentioned doesn’t give the company the kind of marketing data it needs. To solve this problem, more specific attribution models are developed to focus in on whichever part of the sales journey the business is most interested in.

What Do You Do With Attribution Data?

Just to clarify: when we talk about these attribution models and doling out credit within them, it most often entails percentages of credit in their raw form. This credit is then used to influence later marketing decisions. 

For example, if a company is using a first-touch attribution model, they’ll have a percentage for how many of their conversions originated from a given source. Notice how much less effective this approach would be in a sales journey with only one possible source: a mom-and-pop shop with no advertising other than word of mouth and walk-ins doesn’t really need to dole out credit. 

That being said, when there are many potential sources, keeping track with a first-touch attribution model can help the company see which introductory touches are most effective. In a similar vein, a last-touch attribution model can show which of many sources actually leads to revenue most dependably. 

10 Tips For Better Attribution Modeling

Now that we have a basic theory of attribution modeling worked out, here are some tips for making the most of whichever model you choose.

1. Study Your Sales Cycle

It almost goes without saying but you would be surprised how many companies and marketing departments get this one wrong. If you want to choose the best data attribution model, a comprehensive understanding of the sales journey is imperative. 

Any creative familiar with design thinking and personas can tell you how vital it is to employ user research and sales data to establish who your actual users are and what their true experience is like rather than relying on assumptions or what decision-makers want all of these experiences to look like. It’s the same with the sales cycle. 

We know the dream – a commercial so good it captivates and generates sales right away. But we also know the reality, which is that people hear about a brand, then see ads, then hear more, then perhaps engage with branded content, see reviews, see an in-store demo, laugh at a viral commercial, see a sales ad, and etc. etc. ad infinitum until they at long last make a purchase. 

Do the research to find out exactly how people are finding out about your products and services and use that to inform the marketing efforts that you measure with your attribution model. 

2. Touch Points Happen Constantly

This guide has focused mainly on the touch points that happen before a conversion for the sake of illustrating attribution models. It’s also important to understand that touches happen during and after a sale. 

Product reviews are a good example of a touch point that occurs during the purchasing process. Think of any big eCommerce site and how users will compare reviews between products to decide which to buy. From the vantage point of the host site, the sale (i.e. buying something off their platform) has already been made, it’s just the particulars of which item to buy that remain.

Neglect touch points like feedback, upselling emails, or renewal reminders and you could very well start hemorrhaging customers fast. Make sure you have an attribution model that doesn’t stop at the sale or you could be jeopardizing your return customer base. 

3. Establish Benchmark Touch Points

If you’re putting serious cash behind your marketing outreach, you can give credit for upstream touch points for leading to staple touch points. Think of it as a kind of model within a model. 

This is particularly effective if you have a strongly branded marketing method like newsletters, catalogs, or a popular social media account. If you find that getting people to interact with these branded messages strongly correlates with conversions, then you can measure how effective smaller touch points are at directing user engagement to the benchmark touches. 

4. Leave Research Open-Ended

Are you absolutely positive you know every single possible avenue for introductory interactions between potential customers and your brand? Sure, there are some you pay for, but in today’s vast user-focused internet sphere, the company could be getting attention from surprising sources. 

Review sites, social media, and the news are all potential sources of free publicity. As long as links can be shared, there is very likely to be some first-touch potential at the top of your sales funnel that you don’t explicitly create. That’s a great thing. Just make sure your research can discover what are known as ‘missing touch points,’ the known unknowns of the sales cycle.

5. Be Customer-Service Minded

It’s well-established by now that basically every job is a customer service job in the end. Whatever your product or service, the fact that people use it means they’re likely to need some kind of support eventually. 

If your goal is to be a market disruptor, you’re going to have to accept that people will have some confusion with a new product. But you can turn that into a huge net positive with robust customer service options that inform users about the product, answer their questions, and help troubleshoot if the experience is less than ideal. These are post-purchase touch points that have to be included in your attribution modeling. 

6. Allocate For Your Budget

It’s all good and well to measure your sales journey from as many touch points as possible. Just remember that the whole point of these attribution models is to organize how the company puts a percentage of its earnings back into marketing. Using a multi-touch model like the time-decay or U-shaped models where credit is given to every touch point could create an unactionable marketing strategy. 

What we’re trying to say is that you have to be realistic about the marketing budget you’re working with. If you’re just starting out or working within a niche part of a much larger ad strategy, then an attribution model that could potentially put your marketing creatives to work on content for a dozen different touch points is not the best one for your company.

7. Use Data-Driven Attribution

More than just a general category, data-driven attribution (DDA) is the cutting edge of attribution modeling. It includes algorithms and artificial intelligence to find patterns from a huge swath of sources and massive amounts of information, usually more than human eyes could sift through in a reasonable amount of time. 

As technology continues to improve and become more affordable, DDA is poised to become the norm. That’s great news for businesses and the marketing teams responsible for building brands because it allows all the important decision-makers to have an incredibly detailed look at all aspects of the sales cycle right away. 

It also helps remove biases and give a real-world look at how attention is garnered by branded content and what the actual sales cycle looks like as opposed to the one the marketing team thinks they have designed. In many cases, using DDA is the best way to find those missing touch points so you can invest resources in them properly.

8. Collect Data All The Time

User data is as good as gold for UX designers and other creatives concerned with merging business goals and customer desires. Attribution models are no exception. Whether you’re using the AI-based DDA we touched on in our seventh tip or staying classic, you’re going to have to have as much data as possible. 

You can make sure you have the most effective data (rather than just heaps of seemingly random information) by including users with feedback mechanisms and allowing them to use reviews. Focus groups and interviews are also helpful in this regard. 

Most importantly, be transparent if you’re going to collect data. No one will be thrilled to find out they’ve been tracked and in some places you could be in some legal trouble if you don’t have the requisite disclaimers and notifications before you start collecting user data. 

The more you empower users to leave their honest opinions and have a measurable effect on the course of business, the more they’ll appreciate your brand and the better you’ll perform with other users. That starts with collecting data and it’s maintained over the long-term with the right attribution modeling.

Conclusion:

Choosing the right attribution model for your product and business is one of the most important tools for strong, effective marketing. Without understanding how people are entering your sales cycle or which element finally drives them to a conversion, you can’t effectively target your branded material. 

There are plenty of models available. As many companies get more in-depth in their marketing attempts, they find that neither first-touch nor last-touch attribution models cut it. Customized multi-touch modeling used in tandem with DDA and other automated AI tools helps sift through tons of user data and make an informed decision about the future of your marketing efforts. 

Use the information in this guide to select the best possible data attribution model for your company and its particular context. When you know what’s working, you know where your resources need to be invested.