Determining the Right Marketing Attribution Model for Your Hotel
By Stuart Butler Chief Operating Officer, Fuel Travel | January 12, 2020
Can we all agree that measuring the performance of your marketing budget is very important? Without the ability to track your dollars, you're left making uneducated decisions and you're likely wasting a significant amount of money on campaigns or channels that are simply not effective. So, if we all agree that it's important, why are so many people struggling to figure out how to do it, and why can nobody agree on a simple but universal methodology?
There are so many tools available to us, many of which are free to use. The majority of our advertising dollars are now being invested in online campaigns. The majority of our bookings are coming from online channels - and those that still come in over the phone can be tracked fairly easily by most modern call-tracking solutions. On the surface, this seems like a simple nut to crack.
In fact, many vendors in the hospitality industry claim to have it all figured out and will provide you with a beautiful report at the end of each month summarizing the success you've had by using their display networks, their META campaigns, or their fancy widget that helps you drive more direct bookings. The report will confidently tell you how they drove a 20:1 ROI and how great they are at driving new business to your property.
For a lot of hotel marketers, there is a comfort in taking these self-reported numbers at face value. It certainly makes it easier to justify to your boss that you're making a positive impact, but is it accurately depicting the impact that the vendor is having on your bottom line? In most cases, the answer is an emphatic NO!
Here's the thing. The majority of vendors are going to provide you with reports that shine the best possible light on themselves. They have a vested interest in convincing their customers that they are an invaluable component to the overall marketing mix. Trusting your vendor to tell you how well they are performing would be the same as a guest allowing the hotel to leave a review on their behalf. How many hotels would be completely objective? There would be nothing but 5 star reviews from every guest.
Another way to look at it is to consider the relationship between me and my children. I certainly don't allow them to simply tell me what their school grades are for the semester. I let an independent source (i.e. the teachers) provide that information.
It's time the industry stops allowing their vendors to tell them how to track their performance and starts taking back control of the decision-making process. It's time we all started holding all of our investments to the same standards and doing so by using the same attribution models for everything.
There, I've said it. "Attribution model." For many this is a dirty phrase. For most, it's confusing. It means different things to different people. Even the most expert among us would tell you that there's no silver bullet, no one-size-fits-all solution. It's really, really complicated. The reality is that there is no right or wrong. What works for you may not work for someone else, and vice versa.
Regardless, I believe that you should be paying attention. Otherwise, you run the risk of making decisions from inaccurate information. A prime example of this is when it comes to display advertising. I'm not going to call out one specific vendor, because the majority of them follow the same bad practices. Part of the credit they claim may be based on something like the following: 100% attribution for a view-through conversion with a 12-day cookie. Let's break that down.
They are saying that if your display ad was served to a consumer 11 days and 23 hours ago, and they came to your website today and booked, then the ONLY reason they booked was because of that ad that was served. This doesn't take into account what the website was, where the ad was served on that site, whether the person scrolled down to see the ad, whether they consciously or unconsciously saw the ad, whether the ad had an influence on them, whether the guest had stayed with you before, whether the guest had seen an ad for you somewhere else, whether you'd sent them an email, whether they'd performed a google search, whether they visited an OTA, whether they'd read a review, whether they'd read an article on your website, whether their friend had just had a lovely stay at your property …
You get the point, right? Though the ad potentially (and likely) contributed something to the transaction, it is highly unlikely that it was the singular reason that they booked, especially considering all of the things that could have happened in the past 11+ days. The over-stating wouldn't be too bad, but it's impacting where hoteliers invest their money, and in some cases, the compensation that you give to the vendor is directly tied to the revenue that they claim credit for.
I challenge everyone to protect yourself by pushing back against these overly simplistic and overly generous attribution models.The real contribution is somewhere between 0% and 100%. What you'll find is that the percentage is going to depend greatly on many factors and which attribution model you deem to be most relevant and accurate for your business. Below, I will attempt to give you some foundational information so that you can make a more informed decision in the future.
What is an Attribution Model?
An attribution model is a set of rules or a formula that a hotel may use in order to assign all or a portion of the credit for bookings or other conversions to a source. That source may be an advertising campaign, a piece of software, or some other marketing initiatives.
For example, if my hotel generated $1,000 in room bookings, what sources contributed to that revenue and how can I assign a value to that source so that I know which of my efforts is yielding the best results?
Why should hotels use an attribution model?
Attribution modelling creates an apples-to-apples comparison for your marketing efforts and allows hotel marketers to make smart decisions about where they should invest more and where they should pull back.
Marketing has become significantly more complicated over the years and requires more sophisticated tracking and analysis to understand the impact that your marketing dollars are having on your hotel's business.
Examples of Attribution Models
There are many different attribution models out there. There are positives and negatives to each. It's up to you to ensure that you understand the information that the specific model is providing and that the insight you glean from it is always taken in context. There's not necessarily one attribution model that works perfectly for your business. You may find that you can crunch your numbers multiple ways to gain a broader perspective. For example, top of funnel vs. bottom of funnel initiatives may be more accurately compared using different models from each other.
In order to make this as tangible as possible, let's discuss a specific scenario:
A potential guest finds your hotel website by clicking one of your paid search ads in Google. The consumer then returns two weeks later by clicking over from a Facebook organic post that has a campaign ID. That same day, they come back a third time via one of your guest history email special offers, and a few hours later, returns again by typing your url into a browser and making a booking that was worth $1,000.
The following is not an exhaustive list but it covers some of the more commonly used attribution models:
1. Last Click Attribution – the last click or source of traffic that occurred prior to the booking-in this case, the Direct channel-would receive 100% of the credit for the sale.
Direct Booking = $1,000
Opinion: This model over-prioritizes the final click and completely eliminates any interactions that occurred higher up the funnel. Unfortunately, this is a model that is commonly used because it's very simple to implement and easy to understand.
2. Last Campaign Attribution – all direct traffic (and often referred traffic) is ignored, and 100% of the credit for the sale goes to the last channel that the customer clicked through from before converting-in this case, the Email channel.
Guest History Email = $1,000
Opinion: This does a good job of tracking activity of managed campaigns but falls short because it ignores any business you're gaining organically and, similar to the last-click model, it doesn't take multiple touch points into account at all. Where it can be useful is in holding multiple campaigns of the same type to the same standards. For example, if I want to compare all of the emails I've sent out over the past 12 months to determine which subject lines or offers were most effective, this model can help.
3. First Interaction Attribution – the first touchpoint-in this case, the Google Ad campaign-would receive 100% of the credit for the sale.
Google Ads = $1,000
Opinion: This model helps you identify parts of your marketing mix that are early in the funnel and that are often overshadowed by bottom-of-funnel activity. This model helps paint a different picture than the first two models, but it is equally as flawed and should be used only in the appropriate context.
4. Linear Attribution – each touchpoint in the conversion path-in this case the Paid Search, Facebook, Email special, and Direct channels-would share equal credit (25% each) for the sale.
Google Ad = $250
Facebook = $250
Email = $250
Direct = $250
Opinion: This model is a little less discriminating when it comes to where things fall in the funnel, which is why a lot of folks choose a model like this. However, this model assumes that all things are contributing at the same level, regardless of timing, behavior, and the effectiveness of the touchpoint.
5. Time Decay Attribution – the touch points closest in time to the sale or conversion get most of the credit. In this particular sale, the Direct and Email channels would receive the most credit because the customer interacted with them within a few hours of conversion. Facebook would receive less credit than either the Direct or Email channels. Since the Google Ads interaction occurred one week earlier, this channel would receive significantly less credit.
Google Ad = $50
Facebook = $200
Email = $375
Direct = $375
Opinion: We're beginning to get a little more complicated with this model. You would have to determine the rate at which the attribution decay occurs. Is it a percentage each day, or do you group some time periods together? Most people can understand the logic behind the fact that a touchpoint the same day as the conversion may have had a bigger impact that a touchpoint from 30 days ago, but exactly how much more is totally subjective.
6. U-Shaped/Position Based Attribution – this is where first and last interaction is prioritized over everything else. For example: 40% credit may be assigned to the first and last interaction, and the remaining 20% credit is distributed evenly to the middle interactions.
Google Ad = $400
Facebook = $100
Email = $100
Direct = $400
Opinion: This model makes the assumption that the initial interest was triggered by the first event, without which there wouldn't have been a booking, and that the last interaction was what finally pushed the consumer over the edge to a booking. Unfortunately, the model doesn't take recency into account when considering the influence that the middle touchpoints had. Also, there doesn't appear to be a consensus in terms of what percentages should be applied to first and last touch points.
7. L-Shaped Attribution – the first click will get a large percentage of the credit and the remaining touch points will have the remainder spread evenly between them. In this case, the first interaction gets 40% and the remaining get 20% each.
Google Ad = $400
Facebook = $200
Email = $200
Direct = $200
Opinion: The L-shape is a hybrid between a first click and a linear model. The initiator of the transaction, in this case Google Ads, gets the lion's share but credit is also given to the other contributors. Some folks also use a reverse L-shape model, where it's actually the last click that gets the majority. This is really a matter of personal preference and depends entirely on how you personally value the touchpoints.
8. J-Shaped Attribution – similar to a reverse L-shape model, this model gives extra credit to the last interaction. However, the difference here is that the first touch point also gets a greater credit, but not to the same degree as the last interaction. Let's assign the last click a 50% attribution, the first click a 30% and the remaining touch points get an even split of the remainder.
Google Ad = $300
Facebook = $100
Email = $100
Direct = $500
Opinion: Again, this model depends greatly on what you value as most impactful. Some folks will also flip this model and use a reverse J-shape, giving more credit to the first interaction but still giving a bump to the last click.
9. Combination/Customized Attribution – my personal preference is to look at a combination of these models by either combining time-decay with a reverse J-shape, or by simply running multiple models and allowing the story that each one tells to inform your decision making. There's really no limit to how complex you can make the model. You're limited only by the data you're able to collect and the time you have to interpret it.
Factors to Consider When Selecting an Attribution Model
As previously mentioned, there isn't one model that can be considered best. They all do slightly different things and tell slightly different stories. When selecting your model(s), there are many factors that you should keep in mind. Here are a few to get you thinking the right way:
- Is the data accurate and what was the source?
- Do I know the different length of cookies?
- How much time has there been between interaction and conversion?
- Where was the interaction within the funnel?
- Was it new or repeat business?
- What is the lifetime value of the customer?
- Did the consumer use multiple devices?
- Are you including the management cost in your ROI/ROAS calculatons?
- What factors can't you see or track?
- What time of year is it? What promotions am I running? What rate am I distributing to third-parties? etc.
If you're thinking that this is confusing and difficult, you're not alone. Everyone I speak with is struggling with this. There's a reason that there isn't a gold standard for measuring success. Though some folks may have more of it figured out than others, everyone is wrestling with it and trying to determine whether they are doing it right. My advice to you is to make sure you follow these three simple steps:
1. Realize that attribution theft is real
Everyone is going to try to take as much credit as you allow them to get away with. It's your job to poke and prod and question your vendors. Trust but verify by making sure that you understand the model they are using and by also looking at their results through a different model that makes sense to you.
2. Gain a good understanding of attribution modelling
This may not come naturally to everyone, but knowledge is power and you have to at least understand the fundamentals by looking at different models and analyzing the pros and cons of each. Start by performing some basic Google searches and start reading blogs and watching videos. There's a plethora of content and resources available to help you formulate a game plan.
3. Decide on what makes sense for your business and what you and your stakeholders are comfortable with
There's a trade off between how precise you want to be and what information you actually need in order to make a favorable decision. The sooner you realize that it doesn't have to be perfect, the better. You can drive yourself insane trying to track everything perfectly and it won't make a single dollar difference to your bottom line. Just make sure that you hold all your efforts accountable to the same standards.
Sometimes, it helps to look at the data using multiple attribution models to get a better grasp of how specific channels impact your business at different points throughout the funnel. Just don't overthink it.
HotelExecutive retains the copyright to the articles published in the Hotel Business Review. Articles cannot be republished without prior written consent by HotelExecutive.