Over the last 20 years, the world has come to terms with the issues that come with a surplus of data. Primarily, our focus has been on the physical challenge of acquiring and storing data, doing it in a cost-effective way and providing real-time and secured access to users. The world is ‘on it.’
We’ve seen great innovation in this regard, from ‘hyper scale’ data centers to data warehouses and lakes, and now, open data structures, like data lakehouses. These incredible examples of data engineering, however, have only skimmed the surface in terms of the deeper issues companies face when trying to reach an always-on ‘single source of truth,’ per Marc Benioff.
Getting to the heart of unifying data requires a higher-order of intelligence that involves identity, relationships and truth. This may sound like a psychoanalysis, but no. We’re talking about the next generation of challenges big data will need to endure. This is true for all sectors. But how will these next-level problems manifest themselves within B2B?
The identity crisis: Move over, anonymous customer
Companies currently have heightened anxiety over the prospects of a cookieless world. But the identity crisis B2B companies should be concerned about actually has much less to do with tracking anonymous users.
B2B companies shouldn’t simply track the anonymous ID who clicked on a Facebook ad. They need to monitor the many ways they are engaging with people and companies, from partners to vendors and customers. So, what we’re talking about in this case pertains to more fundamental B2B data around company deals, subscriptions, contacts, employees, and more.
Here’s what’s behind B2B’s identity issues:
1. B2B companies are still relying on “traditional,” offline-by-nature identities (physical addresses, phone numbers, contact cards and financial profiles)that need to be aligned with the online world, including social, web and mobile identities.
2. CRMs and ERPs are still the main conduit for B2B data. Those systems, however, tend to produce multiple versions of the same people or companies because they are relying on manual data inputs by multiple departments.
3. To add fuel to the fire, there’s a fast-growing amount of digital channels in which companies interact with their surroundings (chat, email, mobile apps, product logins, Slack channels, Zoom chats, etc.). We’re left with a lot of unwanted Mickey Mouses and an exploding number of ‘loose’ identities. This not only messes with our data, but it stunts the efforts of any enterprise company to consistently engage with people.
4. Finally, B2B identities are much more dynamic than B2C identities. People will change positions and jobs, new companies start and close, acquire, merge, and change at a higher and more frequent pace than ever.
The challenges don’t stop there. The question around ‘what is identity’ will become even more undefined as the Metaverse and other virtual platforms come into play. We already have statistical measures that can solve these issues. In the world of B2B marketing, algorithms are set in place to determine which John Smith is the actual target customer. But as these identities become more complex – less physical and more virtual, with an added layer of what’s real and what’s fake – B2B data will require a higher order of non-deterministic AI algorithms to solve for identity resolution at scale.
Relationships issues: Understanding connections through data
Data is never singular. It’s packed with connections, history and relationships. Just like humans are hardwired to connect with other humans, data science must be robust enough to not only identify those connections but then interpret the meaning for added context.
For instance, knowing that a target customer’s company is owned by a larger company makes a big difference for a B2B sales and marketing team. Understanding connections between and outside of the buying team (e.g. knowing where people last worked or how they are connected with other professionals from previous careers) is invaluable as well. Explicit data about many of those relationships may be hard to acquire, but the good news is that AI algorithms can be trained to detect those relationships based on implicit hints. And, again, once machines are well trained, they can do that on an enormous scale. Data can be limited sometimes, but not the models you build off of it.
Uncovering B2B network relationships can become a huge asset to support data driven strategies. So much so, thought leaders are devoting hours of discussion and complex analysis and science surrounding network analysis.
The truth complex: Use AI to sift through your ‘fake news’ data
Online, especially across our social platforms, we are encountering a constant debate around what is true and authentic. A surplus of information also means an influx of misinformation. And with bots and easy access to blogs and content machines, studies have shown that with information increasing, the quality is decreasing.
We have similar issues in the B2B data world, and we have to train machine learning to decipher the truth. When targeting potential customers, the CRM may identify Amazon as a retail company. The marketing system, however, says it’s a software company. A third system says it’s an IT company because of its AWS business. Multiple versions of one customer may leave you with three salespeople working separately on the same account. So, which is it?
Just like social media platforms are using their best data scientists to detect fake news and unreliable information in real time and on a massive scale, B2B should leverage AI to detect unreliable data that can mislead its operation. AI will be invaluable at pointing out discrepancies and suggesting which is more likely to be true, but in many cases the truth will also be subjective and require human oversight to guide the machine. We see this in the example of deducing if Amazon should be classified as a software company, or not. The right approach will be AI making predictions and humans directing the answers, and then that all feeds back into the machine to completely take over.
The good news is that while B2B data is encountering an entirely new level of challenges that require the deepest “therapy” and attention, AI is constantly undergoing innovation to meet the demands. Our data might be limited, but the applications of AI are limitless. Only AI – paired with human guidance – will be able to solve these worsening issues for the enterprise and make our data, in essence, better.