Bridging the Gap – Multi Channel Attribution Models for Marketers
We marketers LOVE measurement and attribution. Not only is it exciting and gratifying to see the data roll in, it helps make our work feel more tangible and meaningful. Unfortunately, accurate measurement is not always easy (especially when running integrated campaigns across traditional and digital channels). How many online impressions does it take to drive a store visit? How many TV spots do we need to run before we get a website traffic lift? We have gotten pretty good at understanding the online customer journey through multi-touch attribution, but as much as we love cookies and pixels, they just don’t cut it when trying to understand online to offline activity.
The good news? The big players are already developing varying marketing attribution models, and new metrics and platforms are emerging to help solve this riddle. Today I’m going to share some new trends in multi-channel attribution, including online to offline measurement (and vice versa), and I’ll offer suggestions on how to start linking your integrated media campaigns.
Our friends at Google have a solution for us in the form of an Adwords metric called “In-Store Visits”. This data is anonymous and can’t be tied to an individual, but is certainly a step in the right direction in linking online ad interaction with physical store traffic. In essence, this metric tracks how many people click your mobile ad and then visit your brick and mortar store(s), but it is still pretty new so it is not without limitation.
One drawback is this is tied to mobile ad interactions only, so desktop is excluded, and is only (really) working for Paid Search at the moment. Google says it works for YouTube and Display as well, but we haven’t seen those metrics at the same level just yet. In addition, only the larger advertisers with many physical locations are able to even view this metric, which leaves the local flower shop out for now. Here is Google’s eligibility criteria that must be in place before you can take advantage of this opportunity:
- Receive thousands of ad clicks and many store visits
- Have a Google My Business account linked to your AdWords account
- Create each of your store locations in your Google My Business account
- Have at least 90% of your linked locations verified in Google My Business
- Ensure location extensions are active in your account
- Have sufficient store visits data on the backend to attribute to ad click traffic and pass user privacy thresholds
So, this is mostly applicable for national retailers with high media spends and many stores for now. But how does Google measure an In-Store visit? I’m glad you asked! It uses the following criteria:
- Time at the location
- Google Maps searches or navigation to the location
- Google web searches
- The strength and accuracy of location signals (GPS, wifi, etc.)
- Google’s historical accuracy in identifying visitors to your locations
- Other proprietary factors
Google then analyzes this collective data and only counts a visit when the combination of these factors leads them to be confident that there was indeed a visit. In my opinion, this is why the in-store visits look relatively small compared to your paid web traffic and spend for now. Google is doing its best to provide accurate data, and exclude anomalies so that they can improve the measurement tool over time.
We are monitoring this with a few of our retail clients and are starting to see it gain traction; however, due to the current limitations and confidence issue listed above, we see a relatively small percentage of the overall mobile traffic receiving the store visit attribution. That said, analyzing this information even as proxy can provide fascinating insights. Here’s Google’s suggestion on how to assign value to the store visit metric:
[(# of Store Visit Conversions) x (In-Store Conversion Rate) x (Average Order Value)] / Adwords Spend
So, let’s say that our campaigns drove 10,000 store visits, our in-store conversion rate is 40%, the average order value is $75, and we spent $20,000 on the campaign. We would be looking at:
[10,000 x 0.40 x $75) = $300,000 attributable offline revenue / $20,000 in spend = 15 attributable ROAS
Without even looking at the direct online revenue coming in from this campaign, we already have a 15 attributable ROAS to help us gauge success. This is compelling.
Just like customer lifetime value, this analysis and attribution will take a bit of work on the business side to understand in-store conversion rates and average order value. But if we start thinking about our digital campaigns with a physical attribution lens, we can begin to understand the true value and role of those campaigns in the overall path to purchase. Despite the current limitations, savvy marketers will want to keep an eye on this metric when evaluating campaign efficacy.
Facebook has a little ditty called “Offline Conversions” leveraged through “Offline Events” that can be linked to your brick and mortar sales data. This physical data can be uploaded manually into Facebook, or it can be automatically passed through your CRM (limited to Facebook’s current partners in this space for now). This is extremely exciting for online marketing attribution as we will begin to draw direct lines between campaigns and in-store sales (without the need for complex attribution modeling and equations).
Linking the sales data back to your social campaigns is super powerful in and of itself, but what I really love about Facebook’s system is we can remarket complimentary and related products back to the consumer to cross-sell, up-sell, and promote relevant offers! The possibilities here are endless, but I would caution the eager marketer to keep it classy and not annoy people. I would limit the frequency for starters and make sure that only relevant products are being pulled from the product feed for the remarketing ads. That said, if you can dial this in the high relevancy and “B2ME” targeting should provide you with really great returns.
That’s right folks, burgeoning social platform Snapchat is at the table for this one as well. Snapchat has been investing in data partners over the past couple of years, and this is clearly paying dividends. Snapchat’s new “Snap to Store’ system is able to track how many people visit a physical location within 1 week of being exposed to a vertical snap ad.
There are, or course, limitations on this one as well. Snap will only access the location data when people have the app open, and it will aggregate data from large samples of users to place them into categories like ‘restaurants’ and ‘grocery stores’, but not specific stores or venues. This also isn’t available for everyone just yet, but rather just those that “reach a certain spending threshold” aka advertisers with deep pockets.
Unfortunately, we can’t use this to elevate a remarketing strategy (like Facebook), but we can start to think about the visits in the same way that we use the Google In-Store equation. Even just as a proxy, it’s still compelling data to consider along with your ‘swipe up’ rates and ‘website visits’ coming from your snap campaigns. Though still in infancy, Snapchat is 100% mobile, so this kind of attribution potential makes me drool (like an infant).
But Wait, There’s More!
It’s not just the big fish in this pond with their fancy metrics and attribution models, there are many alternatives out there that can make waves in this space as well.
- Call Tracking – Are phone calls/leads a source of revenue for your business? Call tracking systems like Dialogtech provide interesting solutions to attribute phone calls driven by online and offline activity. In essence, call tracking switches your actual phone number for a vanity one, and then connects calls coming to the vanity number to your actual number. This allows us to track the call to the exact location where the caller was exposed. We can use one number for our online ads, another for our TV campaigns, and another on the website and accurately track these calls back to the source.
- Third Party Display & Attribution – There are many new players entering this space right now including Native Touch, GroudTruth, Cluep, and more that have their own proprietary systems for tying online ads to physical foot traffic. Most of these folks have systems that tie into location enabled apps, passive wifi triggers, or their own ‘special’ geo fences, but buyer beware when scoping these. We’ve had a few experiences with media partners that did not live up to the promise and ended up being too good to be true
- Incentivize Store Visits – Provide people with a coupon on social or by email that is only redeemable in store. This might be the oldest trick in the book, but it can still be highly effective for tying your online campaigns to physical traffic.
Let’s Go Old School
Traditional media has been playing this game in reverse for some time now, and what might feel like an old school method is still highly effective. Drumroll please…… unique campaign URLs! By using unique, but easy to remember, URLs on your traditional media campaigns, you can begin to understand which traditional sources are translating into online traffic and sales. Similar to call tracking, you can use unique (but related) URLs on your different traditional mediums to understand the overall impact, and how the channels are playing together in the path to purchase.
So, let’s say I have billboards and TV spots running for my business CK.com. On my billboards the URL is carsen.com/CKcares and on TV it’s carsen.com/CKrocks. I can redirect both of those URLs back to my main website, and through the power of UTM tags, I can understand the where traffic coming in is coming from, and the conversions those sources are driving. Word of caution, make sure those extensions are short and memorable otherwises the consumer won’t remember, which will hurt your performance.
The Missing Link
Understanding multi channel attribution in such a way that links traditional and digital media measurement is widely referred to as the Holy Grail for marketers. Though not yet perfect, it’s great to see the tech sharks and minnows alike taking strides toward holistic attribution through the various attribution models and metrics. These are exciting times for marketers, so keep it locked on this blog for future updates in this space.
How are you measuring online to offline activity? Let’s hear about it below!