In a crowded marketplace full of buzzwords, it’s more important than ever for brands to be equipped with the right questions to ask potential identity resolution and onboarding providers. This will ensure the solution they sign-off on will do what they need it to do, work how they need it to work, and help, rather than hinder, their overall data practice.
In this blog post we will highlight the key questions brands should be asking identity resolution and onboarding providers and why these answers matter.
These questions include:
- Is your graph deterministic or probabilistic?
- Where does your data come from? How recent is it?
- How are you validating the data?
- Do you perform data hygiene on incoming client files?
- What is behind your match rates?
Understanding who and where your customers are across all screens and channels is no easy task. Identity resolution creates an incredibly valuable asset for your organization. Think of an identity graph (or ID graph) as an elaborate jigsaw puzzle where each interconnecting puzzle piece contains an element of information that helps brands accurately identify the customer behind any screen or device. The explosion of devices and touch points, combined with siloed or inaccessible data, has left most organizations with only a fragmented view of their customers, an unfinished puzzle.
Valuable data that comprise these puzzle pieces include everything from names, email addresses, physical addresses, phone numbers and cookies, to social handles, mobile device IDs, demographics and lifestyle traits – regardless of whether or not the user has logged in to reveal their identity, even in an otherwise anonymous setting. Faulty identity resolution (low quality, inaccurate or missing data) is just as bad, if not worse, than none at all.
Identity graphs come in many shapes and sizes. It’s essential for brands to distinguish the real, holistic identity graphs from those that provide only partial linkages. Unfortunately, a single label (ID Graph) is used in the marketplace to represent many variations – with few standing up to the real definition.
Brands need to work with vendors that link full profile identities across all touch-points to reap the benefits of individualized marketing that elevates brands and delivers high marketing ROI. Here are the essential questions to ask when evaluating identity vendors:
Is your graph deterministic or probabilistic?
Deterministic matching is the gold standard. It uses a wide array of data to consistently validate individual identities for the highest achievable accuracy. Probabilistic matching uses probability algorithms and other techniques to put the pieces of the identity puzzle together. It’s more like glorified guesswork and is far less accurate, although tends to yield more volume.
Note what some companies call “deterministic” matching, really isn’t. Some think that if they find a single 1:1 match between two data points that it’s a confirmed deterministic pairing. Others, including Throtle, hold themselves to a much higher standard of quality and accuracy that requires multi-source corroboration of all data before declaring that a match is deterministic.
Where does your data come from? How recent is it?
The building of an accurate identity graph is pure data work. An identity graph is not a licensed consumer file, it is a living, breathing entity that needs to be cared for and cultivated in real-time. Each day, people experience transitions in their lives.
Vendors should have access to a constant stream of multi-sourced consumer data sources (email addresses, MAIDs, cookies, etc.). This allows active validation for both offline and online identities, empowering brands with better offline to online matching without sacrificing reach.
Vendors should also have a rolling exercise to update their sourced data daily, if not in real-time. They should also have policies in place to completely rebuild their graphs on at least quarterly basis to ensure the highest level of accuracy.
How are you validating the data?
As stated above, an identity graph is a living, breathing entity that needs to be cared for, cultivated and validated.
Vendors should have a process in place that ensures the highest level of validation and accuracy. We suggest the minimum standard should be validation from no less than three different, independent sources. Additionally, validation should validate identities consistently and in real-time using personal and persistent data points like an email addresses or device ids.
Do you perform data hygiene on incoming client files?
Identity resolution and data onboarding is a DATA business. If a vendor does not have capabilities to manage, cleanse, analyze, and understand data, they cannot onboard correctly. Technology only enables data to be utilized, distributed, and analyzed efficiently, but if you do not have an excellent data pedigree to start with, none of the technology will repair poor data pairings. Data onboarding platforms that are tech-only solutions leave themselves vulnerable to garbage in/garbage out scenarios with low matching and fill results.
Your vendor needs to have a data hygiene process in place. At the end of the day, onboarding must be data-centric. Brands demand individual and deterministic matching. The only way to accomplish this is by doing a lot of data work up front.
First, an onboarder needs to work on the data being provided. This includes data hygiene, deduplication, NCOA (National Change of Address) procedures, and reviewing for data gaps within emails, postal addresses, or intelligence. Ensuring the brand’s data is standardized and able to be accurately matched is the most crucial first step of the onboarding process.
What is behind your match rates?
The rate at which customer data is matched – the “match rate” – depends heavily on a steady flow of multi-source data. An identity graph provides a constant stream of such data in order to dynamically validate a brand’s prospects or customers as they interact in both offline and online channels. The match rate is vital to the success of personalization efforts.
Data accuracy doesn’t happen by itself. The best identity resolution vendors deploy a variety of “data hygiene” steps such as email validation, de-duplication, National Change of Address (NCOA), etc. to ensure a customer’s data is at the highest level of fidelity PRIOR to match to allow for the highest possible match rate.
As the number of consumer touch points grows, so, too, does the quantity of data collected, causing brands to struggle with obtaining an accurate single view of customers.
Working with a vendor that understand the importance of a holistic, deterministic identity graph can help brands overcome these challenges and can do so in a privacy-safe manner. The key is for accurate data to be top priority. Without accurate identity resolution and activation, brands run the risk of alienating loyal customers and wasting valuable marketing dollars.
The overall impact of individualized marketing driven by timely and accurate customer identity resolution is overwhelmingly positive in both qualitative and quantitative terms. The sooner brands embrace it, the sooner they’ll begin realizing those benefits.