Big Data is the Key to the Future of Multi-Device Marketing
Traditionally, multi-device marketing has focused on in-store marketing and improving ecommerce channels, but there are numerous other applications.
Digital marketers must reach customers across multiple devices. According to Criteo Mobile eCommerce Report, 40% of all online transactions involve multiple devices. Brands that failed to engage customers across different platforms are leaving their digital strategy to chance.
Traditionally, multi-device marketing has focused on in-store marketing and improving ecommerce channels, but there are numerous other applications. Companies that employ digital signage can utilize multi-device data to improve their own engagement strategies.
Unfortunately, optimizing a multi-device marketing campaign is often very difficult. Marketers typically only see one device in their analytics reports, even though the customer usually uses three to engage with them. Big data scientists are developing new solutions to help marketers track customers across different platforms.
Challenges acquiring data across different channels
A recent study from eMarketer found that most marketers are struggling to track their cross-device marketing campaigns. The report found only 30% of marketers used any channel tracking at all. Nearly two-thirds that tried tracking data across channels failed to do so effectively.
Establishing a Consumer Identity is the Key to Collecting Big Data
When analyzing customer behavior across devices, data is highly fragmented. Marketers have a hard time parsing together information to tell a coherent story about an individual customer’s behavior. Finding patterns with multiple customers is even more challenging.
MarketingLand’s Mike Sands states that the only solution is to identify the customer’s identity at the beginning of the engagement process.
“As consumers engage with a brand across devices, they create fragmented identities within each platform. So marketers lack a connected view of the customer journey.
But a single view can be created by deterministically matching a brand’s first-party data across all of these devices and tying it back to the individual customer in a privacy-safe way. When brands can recognize their customers as people, they are able to target ads with precision and offer the right experience in the right context as consumers switch between web, mobile, email and brick-and-mortar stores.
Mastering cross-device identity is the key to delivering better experiences, making the most of marketing budgets and improving measurement for continued optimization. What’s more, marketers will be ready for a future where new devices will open up opportunities to engage with customers in new ways.”
There are a variety of ways that you can identify your customers. Here are the two main approaches to identify the customer and collect data on them:
- Deterministic cross-device matching. Deterministic cross-device matching involves collecting identifying information on your customers at the beginning of the engagement process. This information is usually a customer’s email address. If you collect this information during the first stage of engagement, you can use it to monitor customer behavior in the future, regardless of the devices they use.
- Probabilistic models. Probabilistic models are used to draw patterns between two devices, so you can tell if they appear to belong to the same individual. Probabilistic models track when two devices connect to the same networks, travel together and use similar apps.
There are advantages and drawbacks to each of these models. Deterministic cross-device matching is nearly 100% accurate, since any given email only belongs to a single customer. However, there is no guarantee that you can acquire it at the beginning of the engagement funnel. If you can’t use deterministic cross-device matching to identify the customer at the initial stages of your campaign, you will miss crucial data on the effectiveness of the targeting strategy of your campaign, initial creatives and other critical variables.
Probabilistic models can be used to collect data on customers at any stage of the campaign. However, they aren’t as reliable. Their accuracy is especially limited for local campaigns in small communities, because you may be trying to draw comparisons between multiple customers that use the same WiFi networks and often travel in the same social circles (meaning their devices will often be together, which creates a greater level of uncertainty about your ability to identify a given device or customer).
Your best solution is to use both together. Try to identify a customer using probabilistic matching and confirm your hypothesis with a deterministic cross-device matching model.
Cross-Device Marketing Relies Heavily on Big Data
Data is the lynchpin of any marketing strategy. It is even more important for marketing on multiple devices, since it is difficult to understand customer’s behavior. However, acquiring data is a challenge. Fortunately, there a couple of new models that can make it easier.
You must log in to post a comment.