IN-STORE SALES LINKED BACK TO ONLINE MARKETING
Measuring the impact of online marketing in driving in-store sales.
For a business with high-ticket price sales, lots of stores and products that customers want to get up close to, it was essential to have a way of measuring the impact of online marketing in driving store sales.
The challenge : how to optimise marketing budgets when online CPAs are high.
Running Generic Paid Search campaigns for high-CPC categories was costing a lot of money, but our customer was confident that the investment was driving store revenue. Significant website research and user engagement in the category content indicated these browsers were highly engaged, but tying an accurate view of revenue to the budget was the challenge.
Due to the nature of the products, online sales were not necessarily expected in great volumes, but continued justification of budgets was required. To further complicate things, the long research process most customers followed meant that a multi-touch view of the role of online channels was critical.
Key challenges
- Expensive PPC campaigns for Generics
- Long research process prior to sale
- Limited online conversions
- Attributing store revenue was anecdotal at best
- High online CPAs needed justification
- Multi-touch, multi-device challenge of online journeys
Connecting ‘pre-purchase’ actions online to in-store transactions
We designed and implemented a measurement framework that encompassed multi-touch digital channel analysis, custom on-site event tracking to track key online actions, a solution to join up cross-device usage prior to purchases, and a privacy-friendly anonymous identifier that enabled us to link online journeys to store transactions.
-
Worked with CRM team to define a consistent anonymous identifier, integrate into data feeds and email activity.
-
Designed custom Google Tag Manager triggers and tags for linking key pre-sale events, capture device IDs and anonymous IDs.
-
Data engineering processes to ingest new feeds, capture custom variables, construct a user graph database.
From guess work to a robust attribution framework for store sales.
Every store purchase could now be linked – via the pseudonymous identifier – to ‘pre-sale conversions’ occurring online and the journeys that led to each of these conversions. This allowed us to precisely quantify the number and value – as a headline KPI – of store transactions that had prior online interactions.
There were some limits to the extent to which we could connect ‘every’ store conversion. If a store customer had never converted through one of the key online ‘pre-sale conversions’ through which the anonymous identifier was captured, then their online research would never be linked to the anonymous identifier their store sale was linked to. To workaround this, we worked with the CRM team to implement tracking within emails that would widen the opportunity to connect devices to the anonymous id within the user graph.
Data engineering challenges
There were significant data engineering challenges associated with putting this framework in place.
We had to ingest store transaction data, adapt the way that exposure-to-conversion data was imported from the ad-server to capture various custom variable fields associated with the pre-sale conversion events. The key part of the process was to build and populate a new user graph table which associated anonymous IDs with a device identifier from the online conversion data set. This was then merged with the store transaction data to link the offline conversions to the online journeys.
Measurement & Data Crunching
Marketing Campaign Data
The modules that made
it possible
The following Bright Analytics modules worked together to create this tailored solution:
-
ELT
- Automated data ingestion
- User graph maintenance
- Custom data transformations
-
Semantic Layer
- Online > Store Revenue KPI
- Data harmonisation
- Pre-Sale KPI creation
-
Data Visualisation
- Store dashboards
- Online > Offline Path Analysis
- Automated distribution
FREQUENTLY ASKED QUESTIONS
It will depend on the individual client use case and the information that is available, but the crucial point is that whatever we used it needs to be hashed in such a way as to be unrecognisable. This hashing approach will need to be consistent across all places where this anonymous identifier is being generated.
Each online conversion that a device completes can – usually – be linked to a Cookie ID. This is an anonymous identifier, typically a large number. The key part of the process is the ‘pre-sale’ conversion events where are able to link a Cookie ID to a secondary internal anonymous identifier. With these 2 pieces of information we can connect the offline event to the online events via the shared anonymous internal identifier.
When we are able to see the same anonymous internal identifier across multiple device identifiers, we are able to link and aggregate the online activity associated with the device identifiers to the single internal identifier.
Yes! Once a link is established between the offline conversion and the online device identifiers we can start to examine the historic behaviour and marketing touch-points associated with the online devices. Using GA4 raw data or ad-server data such as CM360 ‘s exposure to conversion data, it’s possible to create a picture of the complete (usual caveats apply) online journey that preceded an offline conversion.
Want to quantify the impact your online marketing is having on store revenues?
Connect with our team today and we can start to explore your use case.