Measuring the relationship between online and offline statistics has been one of the biggest challenges since the founding of the internet. In-store trade is still significant despite the increasing digitalization in recent years.
Being able to precisely measure your offline trade relative to your online trade depends on having the right connection between these two worlds.
Everyone’s heard of this example: a customer searches for a particular product on their phone with Google, Bing or another search engine. An advert for the product is displayed, and the customer clicks on it. In the end, the customer does not buy the product in the online shop, but goes to the physical store in the nearest shopping street: they "convert offline".
Up to now, a variety of analytical and technical simulation models have been used, for example estimating a certain percent of visits to the homepage, or of the most visited pages in the online shop or online platform. Measuring in this way is not only of interest to the biggest online providers, but is also useful for the small Italian restaurant around the corner. For example: the customer searches in the search machine for “Going out for Italian meal”, finds the menu on the page, and then finds out the address. In the last “conversion-step” of the journey, the customer actually goes to the restaurant. This is classic search-behavior, and is something that we all do.
To both optimize and make your Customer Journey measurable, all the phases for a perfect omni-channel symbiosis need to be included and assigned. Whilst measuring clicks on adverts or movements on the landing page are meanwhile classified as a “state of the art” analysis, measuring offline conversions across all devices is a relatively new area. However, it offers enormous potential for the optimization, construction and improvement of your online-marketing campaigns.
Figure 1: Example of a typical customer journey
For some months now, it has been possible to precisely measure the different conversions on Google. However, there are particular requirements for this function, meaning that not everyone can use it, for example:
After further beta-phase testing, this function will most likely be expanded to allow smaller businesses with less budget to measure their offline-conversions. These techniques are also partially used by other providers of online/offline tracking systems.
Measuring online and offline is about making the digital influence on your physical store measurable. You want to build a bridge between the click or interaction online and the conversion or visit to the store offline. Omni-channel strategies will become increasingly important for campaign management and on-page optimization in the future, because people buy or search for products using many different channels.
In general, there are two main challenges involved in the implementation of an omni-channel strategy:
1. The organizational structures of a company (for strategy planning, etc.)
2. The technical possibilities
To be able to precisely measure or evaluate an omni-channel strategy, the customer needs an online-offline authentication. Google approaches this with the following two solutions: the so-called new “In-Store Visits” option and the “User-ID” solution.
With the User ID solution, the customer receives an anonymous ID, such as a log-in in the online shop. By means of the ID, a conversion to the offline store will be recognized, in that, for example, the customer uses the customer card to buy the product.
With the in-store visit, the customer has to be logged in to their Google account on the device with which they clicked on the product. Google Maps also has to be used. If this generates enough data, a prediction can be made. Similar to the Cross Device or other measurements, the data is a combination of what google can measure and a prediction. The calculation is conservative, has a high confidence interval and a very high accuracy level.
How to we reach this type of precise prediction? A range of technical approaches can be used. GPS seems like a good idea at first, but this would only work with detached buildings – it won’t be precise enough in a shopping center or on a shopping street with small stores close together, as it’s important to know the exact boundaries of the store.
Now visitor behavior comes into play. The visitor needs to spend at least 5 minutes in the store, so that the data is gathered from the correct store, rather than one the customer just walks past. The outline of the store, including entrance and exits, will be displayed with customized software. Here, wifi mapping, which measures the strength of public wifi networks in and around buildings, is used. The results from the Google survey-tool will be added to the data and processed.
If your main goal is to bring customers into the physical store, this information (or at least an indication) should ideally be on the homepage of your website. With the new option of In-Store Visits, there are now many ways in which you can guide customers into the physical store. If the customer arrives at the landing page from clicking once on an advert, ideally, you should evaluate this subsequently in your web analysis system, as well as in AdWords with the “In-Store Visits” possibility. Information that the customer has searched for should be displayed to them on the landing page – this is an important factor for success in terms of expanding the customer’s journey into the offline store.
Figure 2: Landing page with all important information needed for a store visit
A call to action in the advert text or on the respective banner is also an ideal way to motivate the customer to click, for example on “Visit now” or “This way to the next store”. This may sound banal, but this step is often forgotten, particularly if the goal of the campaign hasn’t been defined clearly.
After the advert has been placed, as mentioned, the optimization of the landing page is important for the success and measurement of conversions. The customer must be able to find the information that they need quickly and easily on the landing page, ideally without too much scrolling. The page should also load quickly.
Of course, brand-adverts with call to action buttons aren’t the only possible approach – non-brand adverts can also generate store visits. The ROAS omni-channel is generally more successful with brand campaigns, but non-brand adverts also show a high store visit rate. This can vary between 15 and 20% depending on the online shop.
The measuring of mobile is improved significantly when viewed from an omni-channel perspective. It is clear that the nearer a customer comes to an offline store, the higher the store visit rate. I’ll explain this in more detail below.
If the In-Store Visits function is activated in AdWords, a new conversion type appears if you adapt the settings for the in-store visits. Here, you can submit a conversion value for every store visit (among other things). It would be interesting to compare respective shopping baskets from the store and the online shop – you can then try to calculate an omni-channel shopping cart value, which you can submit.
Figure 3: Settings in Google AdWords.
It would be ideal to adapt the CPC-order according to distance and location. As previously mentioned, the probability of an in-store visit increases when the customer is nearby. This means, for example, that I might increase my bid by 30%, meaning there’ll be a higher probability that my adverts will be played to the customer and that they will l therefore turn up in a store. Adapting the bid for the stand location is also a good optimization approach for other applications.
Figure 4: Setting bid adjustment according to location
Figure 5: Setting bid adjustment according to campaign level for mobile devices
You should also consider using a version for mobile devices so that you can connect with users on mobile devices. For example, if a user searches for “Buy second hand Audi A6 Munich” on their smartphone, they clearly show an interest in purchasing, and show a clear interest in this particular product.
To guarantee that your campaign will work in this way, it’s important to correctly construct the campaign and target groups in the AdWords account. Above all, you should implement a segmentation of each campaign that has led to store visits. For example, you can administer your marketing budget more effectively, or book more keywords that would be suited to both an instore visit and pure online conversions.
This is how I would implement my own campaign – with the correct AdGroups for my online shop and another AdGroup suited to in-store visits.
Figure 6: Campaign without in-store visits
Figure 7: Campaign with in-store visits
As previously mentioned, the keywords, aside from the bids regarding location, are deciding. You should test different key word options to find the right combination. Keyword combinations such as “Exact”, “Broad”, and “Phrase” would be fine for a start.
Figure 8: Examples of keywords for an in-store campaign
If the optimization between advert, website and the successful direction of offline-tracking is in tune, there will be nothing standing in the way of an optimal measurement of all omni-channels.
Even if the process of analyzing and evaluating omni-channel measures are still fairly complicated and limited, further interesting approaches to campaign management are being established. It’s fairly certain that these uses of customer-tracking can be integrated in and expanded to the offline world, as well as online tools and platforms.
Currently, the ROI is often driven by e-commerce. Online/offline measurements have the potential to change this. The mobile channel also has to be considered, as there are currently 3-4 more visits over mobile devices than over desktop. If these functions are used correctly, they can be a real game changer.
For a recap, here is an overview of the most important To Dos:
Practice makes perfect!
Analyze your website with Ryte for FREE
Published on 08/25/2017 by Jürgen Eppinger.
After his studies in Media Technology and a few online marketing positions with big firms around Vienna, Jürgen moved to Linz to work with the PPC software agency Smarter Ecommerce. He also teaches in the Marketing sector, is interested in different programming languages of all genres, and is a Raspberry PI enthusiast.