Fairing enables thousands of the fastest growing Shopify stores, from Mack Weldon to Boll & Branch, to better allocate their marketing spend by improving their attribution via customer surveys.

“Half the money I spend on advertising is wasted; the trouble is I don’t know which half.” — John Wanamaker, Marketing Pioneer

The e-commerce revolution removed the requirement for a brand to have a physical location, but as Daniel Gulati of Comcast Ventures so eloquently expressed it, customer acquisition cost (CAC) has now become the digital rent replacing the physical one. Spending marketing dollars on the leading advertising platforms–Google Ads, Facebook, etc.–has become a requirement for growth, since that’s where your customers are spending their time. And since this rent has become a requirement for everyone, the rapidly expanding spend is driving up customer acquisition costs for brands of all sizes.

The key to success for an e-commerce or DTC business is balancing the cost to acquire a customer (CAC) with the customer’s total lifetime value (CLTV). The brands that survive this fierce marketing environment are the ones that accurately attribute marketing dollars to customer revenue and use that insight to properly diversify their marketing spend.

Attribution Is Not Easy

Today, you’re likely almost entirely relying on platform level reporting to track the return on your marketing spend. With this, you’re depending on Google, Facebook, et al. to accurately capture myriad ways your customers discover and ultimately buy your products. But, the customer path to purchase is no longer linear. Consumers today utilize multiple devices to make a transaction, making it harder than ever to understand what drove them to make the purchase. On top of that, newer mediums like podcast advertising and influencer marketing are difficult to measure and performance is oftentimes under reported. Marketers today often rely on last-click attribution, which negates what’s arguably the most important touch-point: where the customer discovered the brand or product. If you solely measure your entire paid marketing performance on self-reported return on ad spend (ROAS) or last-click attribution, you’re only looking at a small part of the picture.

The good news is that behind every transaction you receive there is a customer. For most brands, effectively allocating your marketing spend does not have to be overly complex. By simply talking to your customers, you can build your own attribution model while also building deeper one-on-one customer relationships. Win-Win. This is where attribution surveys come in. At Fairing, we’ve been thinking about this a lot, and we’ve put together this guide to help you understand the pitfalls of relying solely on platform data and how to use attribution surveys to improve your marketing ROI.

Why Self-Reporting Platform Attribution is Bad

First and foremost, platform reporting is a black-box. You’re trusting Google and Facebook to accurately count the impressions and conversions of your marketing campaigns without any simple way to verify those metrics are correct. That’s a pretty great deal for them: they get to tell you how many times your ad was served, how many conversions were attributed to that ad, and how much you have to pay for it. A recent lawsuit argues that Facebook over reported metrics by as much as 900%. 900%!

Second, your customers are now interacting with your brand multiple times on multiple different devices. Accurately tracking a person across all of theses devices is challenging, even for the big companies. This can lead to serious mis-reporting of your campaigns’ performance. Adding dimensions like attribution windows–e.g. 28-day click, 1-day view–makes interpreting these reports even more difficult.

Finally, these platforms are addicting and don’t encourage marketers to develop more innovative solutions. Not to mention, a recent study published by The Correspondent that details digital advertising’s actual lift may be heavily inflated. Many marketers are hesitant to diversify off these self-reporting platforms because measuring success becomes more difficult.

What’s Not Measurable With Digital Tracking

As you start to diversify your advertising budget to channels that aren’t measurable with pixels, you need alternative solutions to measure performance. Tracking non-digital exposure is difficult. Most non-digital media buys–like influencer marketing, out-of-home, or podcast advertising–are written in contracts or utilize platforms where actual impression volume and conversions are difficult to measure.

To measure the success of these campaigns, most brands use discount codes (like “podcast”) or special URLs (like “example.com/subway”) to track the efficacy of offline marketing. Making sure that this attribution information is correctly captured in analytics platforms like Google Analytics can be time-consuming and error-prone, and usually requires some level of technical expertise. Not to mention, not all customers will use these codes or landing pages so the actual campaign performance is often underreported.

These problems are only going to get more prevalent as digital CPMs continue to increase. Larger brands and corporations can introduce media mix modeling into their analysis to understand their offline campaign performance –– unfortunately accuracy with this is difficult to achieve until you have 7-figure monthly budgets.

If we can’t rely on coupons, landing pages, or simple media mix models, how do we measure success?

What is an Attribution Survey?

An attribution survey is simply a survey that helps you identify how your customers discovered your brand.

Often times this survey presents itself in the form of a simple question,“How did you hear about us?”, with a list of responses. This time-tested strategy is not necessarily new or novel; you’ve most likely answered one at some time or another. What is new is your ability to seamlessly build it into your existing customer journey and extract immense learnings from the results. The responses collected via an attribution survey are a form of qualitative data, which when analyzed correctly empowers you to uncover a narrative that quantitative data can't and will never tell.

“The thing I have noticed is when the anecdotes and the data disagree, the anecdotes are usually right. There’s something wrong with the way you are measuring it.” – Jeff Bezos, Amazon CEO

Who Needs An Attribution Survey?

Attribution surveys are helpful for anyone who wants to optimize their marketing spend, unlock new insights, and scale their revenue. They become even more powerful when you start to diversify their media spend away from digital channels and onto mediums like podcasts, billboards, and out-of-home. At Fairing, we’re working with brands on the Shopify ecommerce platform to integrate attribution surveys, but these ideas apply to any vertical that involves acquiring new customers.

To note, attribution surveys don’t work very well for legacy brands. Asking someone how they heard of Coca Cola is obviously not going to be super effective. In this instance, the legacy brand can ask more qualifying questions like “What brought you in today?”.

When should you implement an attribution survey? The short answer is as-soon-as-possible. Pre-launch, if possible. Understanding early where your customers are coming from will help immensely as you start to scale up your marketing spend. You’ll most likely uncover channels you never knew existed.

Implementing Your Attribution Survey

Our attribution survey solution, Fairing, takes less than 5 minutes to get live and is accessible via the Shopify App Store. If you’re not on Shopify, feel free to email us at hello@Fairing.com and we can discuss a custom implementation.

What Question Should You Ask?

Fairing was initially built around the question “How did you hear about us?”

When setting up your attribution survey, there are a few variations of this question you can implement. Currently, over 85% of our customers are using our default question, which is simply “One more thing, how did you hear about us?”. Other versions include “How did you first hear about us?”, “Where did you first discover us?”. They’re all built to collect the same piece of data, so it’s really up to you how you want to phrase it. We rarely see any large variance in survey completion rates for these different types of questions.

The possible list of responses to your survey question should reflect the channels that you’re currently running marketing campaigns, as well as channels that you’ve run in the recent past. A user may have heard about your brand on a specific podcast 4 months ago and finally decided to purchase today–it’s important to capture that. We’ve also found that providing an “Other” option where users can write in their own answers can provide countless insights into channels you never knew existed.

Placement of Your Survey

If you’re a brand selling online, we recommend implementing your survey immediately after someone makes their purchase. This way you can tie back that user’s order information to the survey response for a richer customer and attribution profile. Although you’d might gain some interesting insights from polling all users, the greatest value will be getting responses from customers who actually made a purchase.

Fairing’s survey is inserted directly on the order confirmation page, immediately after a user makes a purchase. This real estate is highly engaging, as users who have placed their order are eager to see a confirmation that their order was actually placed. By placing our survey here, we see survey results as high as 60% and often times even higher. If a user doesn’t complete their survey the first-time, we also place it on the order status page. Shopify, by default, uses this page instead of UPS/FedEx tracking pages so engagement is typically very high.

Survey Completion Rates

As mentioned above, Fairing see survey completion rates upwards of 60%. Achieving statistical significance is difficult and totally dependent on the effect size. Given your marketing initiatives vary as time progresses, it’s best to keep your survey running at all times.

Analyzing and Making Sense Of Your Attribution Survey Data

The issue with most survey products is that they do a great job of collecting survey data and a horrible job at analyzing it. At Fairing, we’re building tools where we know the question being asked, which allows us to build robust reporting around your response data.

When you first launch your attribution survey, you’ll most likely be enamored reading each individual responses. If you haven’t been collecting qualitative data in the past, you may feel a sense of connection with your customers that didn’t exist before, as your customers are now giving you 1-on-1 feedback. If you’ve allowed an “Other” response, you’ll be amazed by the unknown sources of brand discovery. The “Other” responses can also help you optimize the list of response options you provide in your survey. Although time consuming to analyze at scale, dissecting your other responses is crucial to building a full map of where your customers are hearing about your brand. You’ll want to assign your other responses to a category for easier analysis in the future.

A Simple Breakdown of Responses

As your data comes in, you’ll start to see a clear breakdown of responses after roughly 100 orders. This breakdown is your top-of-funnel channel discovery mix. Hopefully the distributions won’t be too much of a surprise, but you most likely will glean some learnings you weren’t expecting.

Individual Channel Analysis

The next step is to analyze categorical responses. When analyzing against an ad platform that reports a conversion number, you can start to build a model around how much they’re over or under reporting. For example, simply compare the number of users who responded with Facebook Ad with the conversions that Facebook is reporting to understand your variance. After doing this for each channel you’ll have a clear representation of how survey response data compares to pixel data. Typically we see Facebook over-reporting prospecting campaigns and all other channels under reporting.

In a traditional attribution model, your survey response for the question “How did you hear about us?” can be closely correlated with first-click. This is especially true for new customers. By using referral source or UTM parameters (two metrics Fairing captures), you can start to paint a clear picture of where your customer heard about your brand and what channel drove them to a conversion. By simplifying the path-to-purchase to two data points (survey response and last-click), you can draw considerable conclusions and build a simplified version of a position-based attribution model which weighs first and last click highest.

For channels that don’t have self-reporting attribution, like podcasting, you’ll want to compare your survey response data with conversions containing both coupon codes and your pre-designated landing pages. You can triangulate these three data points to properly understand the success of a given campaign. Given that not everyone remembers a specified coupon code or landing page, we often see the attribution survey results indicating a higher-than-reported conversion rate for non-digital marketing.

Analyzing Trends in Your Response Data

As your marketing mix changes, so will the distribution of your responses. It’s therefore important to analyze your response data over time. By solely looking at your data in the aggregate you’re not taking into consideration any variances that may be happening in the short-term that you can then capitalize on. For example, if the response Online Blog is suddenly spiking, a journalist may have published a post about your product without directly linking to your site. Typically you can resolve this by analyzing your other responses, finding a mention of the post, and reach out to the author to get a direct link.

Pivoting On Other Dimensions (AOV, LTV, Total Volume, etc)

Other than simply looking at your channel mix, you’re going to want to dive into the details. For example, you can answer questions like what channels provide the highest ROI or customer lifetime value (LTV) or what products people purchase who discover your brand via a subway ad. These analyses provide invaluable insight as you and your team build mental models around allocating your marketing spend. As mentioned earlier, the goal is to maximize the ratio between CAC and CLTV, by using survey data to do so you’ll be one step ahead of your competitors.

Using Attribution Data in 3rd-Parties

To extract even more value out of your attribution survey data, you can utilize it in third-party applications to make analysis and segmentation even more intelligent.

Email Service Provider (ESP)

At the moment, Fairing integrates directly with Klaviyo, allowing you to attach response data to customer profiles. You can then utilize this data to send targeted email campaigns to users based on where they discovered your brand. For example, if a user says they heard about you through the The Daily or the Joe Rogan podcast, you can send them personalized campaigns based on that demographic.

Data Warehouse

If you’re more advanced and are utilizing a data warehouse or working with a company like Daasity, you can push survey response data to your database and have an analyst extract meaningful learnings by comparing response data to everything you already know about your customer (products purchased, location, etc). This exercise helps you create a unified view of your customer and build more personalized campaigns.

Launching Your Attribution Survey

If you’re selling on the Shopify platform, head over to the Shopify App Store and install Fairing. The setup process takes no more than 5 minutes and we provide a free, no questions asked, 14-day free trial.

‍