An interview with Anudit Vikram, SVP, Advanced Marketing Solutions, Dun & Bradstreet
Kwanzoo: Thank you for your time Anudit. Let’s start off with an easy question. What is your role at Dun & Bradstreet?
Vikram: Sure, my role is to drive the product technology and data sciences behind our digital efforts. Within our Sales and Marketing line of business we have been building out our digital products over the last few years and that is what I drive. My primary focus is digitizing our data assets and building products for digital marketing, online advertising, or to use in the adtech ecosystem.
Kwanzoo: The data provider and data management vendor space is getting more crowded every day. How do you view and segment the different players out there?
Vikram: From the B2B perspective, the act of creating a data asset is governed by the birth of the data—“where is the data coming from?” Most third-party data providers out there are scraping websites or tagging pages or building relationships with Demand Side Platforms (DSPs) and other media vendors to get “the firehose” of transaction data coming through. They then build some models around that data and do some analysis to figure out what data assets can be created. At Dun & Bradstreet we talk of our data as being “inherently deterministic”—meaning that the data that we collect is provided directly from companies themselves or through transaction systems that the companies are using. This data is explicitly mentioned as such. When a specific company gives us data around revenue we do not run models to determine the revenue, the company has told us that revenue information and we have other native validation points that we use.
Within the B2B space, we see deterministic versus probabilistic data providers. Then the data offerings can be further classified as firmographic, demographic, intent data, and so on. The space is getting “fuzzier” as more providers claim that they offer a wide variety of data.
Kwanzoo: With sales and marketing teams looking to data to be more effective, what is D&B doing to stand out and maintain its leadership position?
Vikram: First, we believe it starts with the underlying quality of the data itself. We feel that we are in a much stronger position regarding that than anyone else. Secondly, a key factor is the scale of our data. We are seeing in the B2B space that data scale is rarely available. It is different for us. For example, we have 99+% coverage of all the businesses in the U.S. market. Today we have data sets and attributes on 24.4 million marketable businesses—and that is pretty much the U.S. market itself. Globally we have about 265+ million businesses that we track and have information on. So, we certainly have data at scale.
Thirdly, we are strong at converting offline data to online assets. Typically, the drop off rate is high during this conversion due to the difficulty in finding cookies, device IDs, and other online identifiers to associate back to companies or individuals. Of those 24+ million companies that we have offline data on, we can find online data for about 16 million of them. What that translates to for the B2B marketer is that in a U.S world of about 70 million business professionals (contacts), he or she can reach about 42 million of them with D&B online data.
Kwanzoo: According to 2016 D&B programmatic research, “65% of B2B marketers are buying advertising programmatically, and 70% are using or plan to use account-based marketing (ABM) in 2017.” Does D&B see ABM as an important market?
Vikram: Yes, we see “account-based marketing” as a label identifying a hot button space within the overall B2B market. In regard to ABM, since D&B is purely in the data business, we feel that we are uniquely positioned to help account-based marketing programs become much more effective and efficient. We can provide the data to help you become hyper-targeted and hyper-focused or identify exactly what you want to go after. You may use the commercial intelligence and the attributes we have on businesses through the D-U-N-S® Number (Data Universal Numbering System) or you may identify specific contacts within businesses that you might want to reach with a specific message. However, we are not an ABM campaign provider like Kwanzoo.
Keep in mind, in the B2B space, we have been supporting strategies such as account-based marketing for quite a while. We’ve been identifying businesses, accounts, and contacts to help our customers have targets for brochures, events, webinars, or some other lead generation effort. It seems like we have been doing this forever. Only now there is new technology that helps package these activities into something a little bit different under a new moniker—account-based marketing.
Kwanzoo: Yes, D&B is striving to be best in class for data in the same way that Kwanzoo is looking to be best in class for ABM display and retargeting. While we agree that many marketers have been doing a variation of ABM for years, we are finding that many have a hard time converting their attention from leads to accounts and from form fills to engagement.
Vikram: I agree. We are seeing the same thing. We do believe that ABM is so much more interesting now because technology has the capability of enhanced cross-channel and multi-channel marketing. That is exciting as no enterprise ABM program can be fully successful through just one channel. Unfortunately, those marketers who have been focused on lead generation typically look to leads and form fills. It is harder for ABM marketers to assign key performance indicators (KPIs) and measure performance. Attribution and reporting across those channels is not that straightforward. It gets more complicated as true B2B marketing requires both online and offline activities to drive opportunities and revenue. We believe that marketers need a universal identifier that can be used to track online and offline attribution. To us, it is the D-U-N-S® Number.
Kwanzoo: You’re right. The holy grail of B2B seems to be able to identify every company and individual business user with a unique persistent identifier across online and offline channels. Can you go into more detail about D&B’s thinking around that?
Vikram: We believe that a unique persistent identifier is of paramount importance. You figure that an average enterprise, B2B deal spans about seven departments in an organization and involves 14 to 17 people. Those people are interacting with you via website, email, events, and phone. The close rarely happens at an online stage in the process. Being able to analyze and measure a campaign with all of its touchpoints and contacts can be cumbersome. You need to pull all the contacts that came from the same company. You need to measure the effectiveness of email, direct mail, events, phone calls, social media, and more. You need a token or identifier that will cover all of those channels. We think that the D-U-N-S® Number best fits that need.
If you use D&B data, all of our cookie information ties back to D-U-N-S® Number. We can tell you which D-U-N-S® Number visited your website, which ones opened your email campaign, etc. As long as you can tie every interaction with somebody to a D-U-N-S®, you are essentially tying it back to a company. So, you can track interactions with 15 different people from the same company as interaction with one company 15 times versus interaction with 15 different people. That difference gives you the insight you need as you review cross-channel marketing reports and data. It doesn’t matter who you are buying your data from or what attributes you use, we can show you which D-U-N-S® interacted. So, if 1,000 people saw your ad and 100 clicked through, we can show that those 100 came from 10 companies. These are the companies that are truly interested in you.
Kwanzoo: So, does this D-U-N-S® identifier approach work across all channels including work addresses, home addresses, and cookie-based and IP-based formats?
Vikram: There is a bit of a caveat. Like most online items, it does not work 100% of the time. We can translate IP addresses back to D-U-N-S®. As I mentioned earlier, we also have cookies and device IDs that are tied back to D-U-N-S®. So, in most cases, if you come online at work, we can look at that work IP address and tie you back to a D-U-N-S®. If you go home and log in, we will most likely have a cookie that will tie back to your company. What makes us unique is that we tie the cookie back to the D-U-N-S® and we do not give the D-U-N-S® to any other data provider.
Kwanzoo: Lookalike modeling is a well-known strategy to find additional targets in B2C, does it also work in B2B such in areas as expanding target account lists? What is D&B doing in this area?
Vikram: Yes, we have found that lookalike modeling does work in B2B but the model is only as good as the quality and quantity of data that goes into the model. More often than not, B2B models tend to rely on smaller data sets. That does not hinder us. With our “strategic profiling” product, we can take a customer list and see which customers are working well for you—and not working well for you. We then overlay that group as a suppression file against our known business universe and will come up with “profiles” (connection of companies) that look like the customer list you gave us.
The reason why we come up with so many profiles is so that you can gain insights from a variety of different perspectives. Profile A may be companies that are similar to your list in customer lifetime value. These companies may have a high lifetime value but be few in number. Profile B may have a lower lifetime value but has more companies in it. You then have the option of creating a very focused campaign to Profile A where each account will give you more money or a broader campaign to Profile B where each account will give you less money but you will have more accounts. Or you could do both. By default, we provide five profiles with our strategic profiling product so most of our customers apply different marketing tactics to each profile. Currently we offer these profiles at the account level. We are investigating various privacy issues and concerns to see if it makes sense for us to also offer the profiles at the contact level.
Kwanzoo: There are more and more predictive tools coming out to help with account selection in ABM? How does D&B view the whole predictive space?
Vikram: We see the predictive tools space as an interesting one. There are many companies coming out with machine learning models, specialized models for doing predictive analytics and lead scoring, and so on. The theory is fantastic. The promise is fabulous. However; I don’t think that anyone has “cracked the code” for providing a scalable and viable product. These predictive data models are only as good as the quality and quantity of data put into them. I am sure that the smart people working on them will only get better. I truly believe in the concept but I feel that we are a long way from being mature yet in that space. On a scale of 1 – 10, we are barely touching a two or three.
Kwanzoo: How do you envision data products and data usage evolving over the next one to two years in the world of B2B and ABM?
Vikram: I spend a lot of time thinking about that question. I truly believe that data is the fuel that drives campaign efficiency, marketing ethicacy, and so on. The way that data will become more recognizable is when you are able to attach more value to it. That will happen when data providers are able to take a data asset and translate it into something that actually returns dollars and cents. Today, in most cases, data assets are two steps removed from the end result we want to achieve. It is only when the value can be tied directly to the data that the data will become more prevalent.
Kwanzoo: Thank you for your time today Anudit. We have just one more question. We are hearing that ABM customers are having a hard time distinguishing data provider products and services as every data provider is promising data scalability, data quality, etc. How do you think prospects can differentiate one data provider from another?
Vikram: Yes, that is a problem. Data is usually at least one step removed from the end result. Let’s say you run a display advertising campaign and you do not get good results. The first inclination is to blame the data as being bad. In reality the campaign could have been bad. You targeted the wrong sites, you were not spending enough money for the media (which can get expensive quickly), etc. It’s tough. Unfortunately, no matter what providers say, the only way to ascertain the value is to test the data through display advertising campaigns, etc.