datamaster summit 2020

Tamr Customer Mastering for High Tech Manufacturing—Demo Days

 

Mingo Sanchez

Senior Sales Engineer

Gain unparalleled insights in your data to power customer 360. Learn how you can break down data silos to achieve high quality, consistent customer data in order to improve regulatory operations, lower their risk and enhance customer experience.

Master customer data across direct and indirect channels and manage customer hierarchies, identify cross- and up-sell opportunities, drive revenue and operational efficiencies.

Transcript

Tamr Customer Mastering for High Tech Manufacturing—Demo Days

Length: 14min 48sec

Tags: Big data, Data, Manufacturing, Demo, Customer, Account manager, Customer relationship management, Account, business, enrichment, Data Universal Numbering System, Information, golden record, Analytics, Different Recordings, Feedback, record, Tamara, operational systems, business decisions

1 Mingo Sanchez on how Tamr can solve your data challenges

Tags: Data, Big data, Manufacturing, operational systems, Industry, Analytics, Demo, Customer, decision, Information, quality, Domingo Sánchez, Organization

00:01 – 00:20

Mingo Sanchez

Hello, everyone, and thank you so much for joining today’s manufacturing demo day presented by Tamr. My name is Mingo Sanchez and I’m a senior sales engineer with Tamr. Today, we’ll be walking through a number of different use cases that are relevant to the manufacturing space and really diving into one in particular. So really looking forward to talking about how Tamr can help solve your data challenges.

00:22 – 00:58

Mingo Sanchez

Now at Tamr, we’ve worked with numerous organizations, and you’d perhaps be surprised at how many of them tell us that big data is just inherently challenging and it’s impossible to work with. Well, at Tamr, what we’ve realized is that big data isn’t the problem. Bad data is you’re not going to be able to make any sort of decisions about your business without having clean, accurate and trusted data to use for those downstream decisions. Because of that, it’s really important to get your data quality in order so that you’re able to make those decisions as effectively as possible, whether that’s having a single view of your customers.

00:58 – 01:23

Mingo Sanchez

customers. you know who you’re marketing to and how you should be marketing to them. Having trusted data to power your analytics so that when you’re making those day to day decisions, you’re actually making the decisions with accurate information or even something as simple as having trustworthy data to use in your operational systems. These are all problems that are really challenging across all industries and particularly when you get into the complexities of the manufacturing industry.

01:26 – 02:00

Mingo Sanchez

Now at Tamr, we know that the legacy approaches of spending months or years building up these complex rules and having highly technical users maintaining those rules simply doesn’t scale with your data needs and challenges. So because of that, we take a very different approach, known as a data ops approach where you’re continuously using your data as an asset. Having those people who understand the data best, providing that feedback and constantly making it better and better so that you always have trustworthy up to date and accurate information will be diving into this more. Once we get into the demo.

2 Why companies choose Tamr

Tags: Manufacturing, Industry, Customer, Machine learning, decision, product

02:03 – 02:36

Mingo Sanchez

Now, when it comes to why companies choose Tamr, there are really three fundamental reasons that stand out above the rest. Driving growth and increasing revenue of their business, reducing the risks associated with the projects that they’re working on, day in and day out and ultimately lowering the total cost of ownership for their architecture. Because Tamr uses a human guided machine learning approach, we’re able to vastly speed up those processing pipelines and ultimately save customers on average 90 percent plus reduction in manual effort in order to clean and curate their data.

02:38 – 03:20

Mingo Sanchez

Now, there are so many different use cases in the manufacturing space that our customers deal with every single day. Everything from cleaning up your list of suppliers so you know who you’re doing business with, to mastering and classifying the parts and products that you’re selling and dealing with, so that you’re able to get that most up to date and accurate information about those parts and products and optimizing your spend so that when you’re making decisions about where to purchase from or who to purchase from, you have the ability to make the best decision possible to save costs as much as possible. But in today’s demonstration, we’re really going to be focusing in on one fundamental use case that’s true across all industries, particularly in manufacturing, which is customer mastering.

3 Demo, Shield Inc. and Forrester

Tags: Account manager, business decisions, Customer, Demo, Forrester, Shield Inc., Shield Inc, Analytics

03:22 – 04:08

Mingo Sanchez

Now, before we jump into the demo itself, I just very quickly wanted to mention that Forrester recently came out with a study where they show that on average customers see a six hundred and forty three percent return on investment with Tamr. So if you’re interested in learning more about this study and how customers are saving money with Tamr, I strongly encourage you to visit our website, Tamr.com to download that study for yourself. And now diving into the demo, we will be going through a customer mastering use case through the eyes of an account manager at a semiconductor company, Shield Inc. Now, Shield Inc has grown over the past few years through mergers and acquisitions, and what they ultimately want to be able to do is get that accurate and trusted view of their customers so that they’ll be able to more effectively cross-sell and upsell to those customers.

04:09 – 04:43

Mingo Sanchez

In today’s demo, it will really be focusing on three key areas. First. We saw systems that our account managers are using day in and day out when they’re doing their jobs after showing how Tamr can be used to make those systems better or peel back the curtain a little bit and show how Tamr can actually be used in order to clean up this data. And finally, we’ll be walking through some example analytics in those downstream systems to show how once you use Tamr to clean up that data, you can use it much more effectively to power business decisions downstream.

4 What is the state of our CRM prior to using Tamr?

Tags: Customer relationship management, Account manager, Customer, Information, Data, Account, enrichment, view, visibility, quality, Feedback, Denbigh Duns

04:44 – 05:26

Mingo Sanchez

Let’s dive in. So starting off here in our CRM application, Salesforce, we can see that we’re looking at an account Lenovo website now I, as the account manager for Lenovo website, know that this is an active customer that we already have an active relationship with. But looking in Salesforce, I can see that there’s a lot of information missing from this account. For example, we can’t even see that indicator that this is an active customer, which I know to be true. We have no visibility into the products that they’re purchasing from us or how much we’re discounting on average. And if I scroll down here, you can see that even as fundamental information as where their office is located isn’t fully populated.

05:28 – 05:50

Mingo Sanchez

Now switching into this account hierarchy view, we can see a number of different accounts, all associated with Lenovo. And while some of these are indeed distinct from one another, I as a person can tell that a few things look like they’re actually duplicates of one another. So they’re slight variance in how these accounts are named, how these addresses are formatted. But I, as a person, can tell if these are actually the same thing as one another.

05:50 – 06:47

Mingo Sanchez

So already I’m starting to question the quality of the data that we have in Salesforce. Now that’s the state of our CRM prior to using Tamr. But once we provide that feedback and are able to clean up the systems, we’re able to get a much more complete and accurate view. So switching into this tab to show the view after Tamr, we can see that once we’d gone through that data unification process with Tamr, we have much more complete and up to date information. Notice how we’re able to get external enrichment information, such as the D&B DUNS number. We’re able to see that this is indeed an active customer that is purchasing products with us and we have those products listed below. And we even have important information like the average discount rate that we’re giving to this customer so that when we do business with them in the future, we’re not going to risk giving them unnecessary discounts and having that price leakage that could negatively affect our business.

5 How does Tamr work?

Tags: Different Recordings, golden record, Information, Tamara, Data Universal Numbering System, Logic, Feedback, Field, enrichment, Account, Machine learning, cluster, more information, SAP, address, pattern, Organization, Rule, Customer

06:49 – 07:32

Mingo Sanchez

Now, obviously, bringing together all this information is really fantastic. But one thing in particular that I want to draw your attention to is up here at the top and thus we have this new field called the tamer ID. And that’s really part of the magic of Tamr is that with just a little bit of work from those users providing feedback on the data, we’re able to join together all the different records associated with this account and ultimately associate those source records with a unique idea that persists over time so that as we bring in new information, we’re able to link that all back to this account. And we’ve talked a lot about how Tamr can clean up these source systems. And without any further ado, let’s dive into how it actually works and tamer.

07:34 – 08:17

Mingo Sanchez

So now you can see that we’re in the tamer user interface, and here we have a golden record, a canonical master record for that same account that we were just viewing in Salesforce. So now we have that same tamer ID linking together.. All this information assisted with that account. In essence, all the information that we’re processing in Tamr is going directly back into those upstream systems so that you always have the most up to date and accurate information possible. And that’s how we have much more complete information than we had in that original record in Salesforce. So we have a lot more information about the address. We have that external enrichment data providing us with that D&B DUNS number, as well as information about the parent organization of this company.

08:17 – 08:56

Mingo Sanchez

So really a much more comprehensive view, but this information didn’t all come from one place. Because this is a golden record that we’ve created in Tamr. We’re actually linking together many different records from our upstream systems in order to get this complete view of Lenovo. So bringing information from systems like Marketo and Salesforce and SAP, you can see that from system to system, we have a lot of variance for how this account is represented, but that’s not a problem for Tamr. We’re ultimately able to group together all of those records and consolidate them into the most trusted record to use for your downstream purposes.

08:59 – 09:40

Mingo Sanchez

Now, how do we actually get to this point where we’re able to link together all those records? Well, that’s where the human and the machine learning come into play. So on this page, on the right hand side, you can see many different records coming from all of these different source systems that we’ve connected to. And you’ll notice that these systems don’t always have consistent formats of data, nor do they have the same fields populated from one record to another. So it’s really difficult to bring together all these different systems and try to write a rule so that that’s able to master those records effectively and accurately, especially if you’re bringing in more systems constantly.

09:41 – 10:02

Mingo Sanchez

But for Tamr, this isn’t a problem because ultimately users are providing that feedback on how to match those records together. And the team was able to jump from this messiness that you see here on the right to these groups of records that you see here on the left. So we call those a cluster of records that Tamr has identified completely on its own as belonging to the same customer.

10:03 – 10:42

Mingo Sanchez

So in this case, we’ve clicked into the cluster for Lenovo, that same account we were showing in the golden record page. And you can see that from system to system, from record to record, we have different ways that these ideas are formatted, so we can’t do something simple, like a drawing on an ID. We have different ways that these names are represented, as well as different representations of these address fields, including the city and state, if they’re even present at all. So I would challenge any rules based system to effectively match together all these records without having a person go in and manually find every single match, which can be a very time intensive and error prone process.

10:43 – 11:12

Mingo Sanchez

So how was Tamr able to match together these different records? Well, that’s where that human guided machine learning comes into play. With Tamr, we don’t require users to be highly technical and write tons of code, all they need to do is be able to fundamentally answer those questions about whether or not records should be matched together and as they provide that feedback to the system. Tamr gets smarter and smarter at picking up on those types of patterns and is able to apply all of that same logic that a person would at machine scale.

11:14 – 12:09

Mingo Sanchez

So the way that we do this in Tamr is by presenting users with pairs of records at a time and having them fundamentally answer those questions about you have record A and record B. Are they the same or are they different? Yes or no? And that’s all you need to do. And as you answer more and more of these questions, Tamr, gets smarter and smarter at handling similar types of records going forward. So when I answer a question saying that these two records should be matched, I’m not just telling Tamr what to do with these two specific records. One said, Tamr is looking at all of these records, all of these points of feedback in aggregate and generalizing those patterns. Learning which fields are important. Similar to those different fields need to be. What do I do in certain fields are missing or an accurate and so on and so forth? So you don’t need to explicitly write out all that logic. Tamr, smart enough to look at what the people are telling it and then answer those questions at scale going forward.

6 Operational systems

Tags: business decisions, operational systems, Analytics, cluster, Information, more information

12:14 – 13:08

Mingo Sanchez

Now, once users have answered these questions and typically it’s a very fast process to get up and running. What do we do next? Well, as I mentioned before, Tamr is able to create those clusters and assign that unique idea that persists over time. So one option is that because you have that full lineage, you can go back into your upstream systems like Salesforce and clean up those values so that you have the most trusted and accurate information in those upstream systems. But not only container clean up those operational systems that you’re using every single day. It can also be used to fit downstream analytic purposes. So prior to using something like Tamr, we didn’t have visibility into how to make those business decisions effectively because we didn’t have trusted information about our accounts. But now that we’ve gone through that data cleansing process with Tamr, you can see that we have a lot more information and can have a much more accurate analytics.

13:09 – 14:00

Mingo Sanchez

So notice how we’re able to get a much more complete hierarchical view of our Lenovo account so that while we have up to this point focused on the website site, we now have visibility into this other related sites that may have similar purchasing patterns. So you can see that we’re selling a lot to Lenovo website right now, but we aren’t selling quite as much at these other sites. But now that I have this information linking together all these different records associated with Lenovo, I’m able to take that pattern for success and replicate across these different sites, potentially getting a lot of upsell opportunities along the way. Now this is just one example analytic, but because you have clean, accurate and up to date information, now that you’ve used Tamr, you’re going to be able to make much better decisions about anything that your business is concerned with going forward.

14:02 – 14:46

Mingo Sanchez

So in conclusion, we’ve shown three key things in today’s demo. We’ve shown how Tamr can be used to clean those upstream operational systems like Salesforce. We’ve shown a little bit about how that process actually works in Tamr, where users are providing that feedback and making those systems better. And then finally, we’ve shown some examples of how with just a little bit of work up front in tamer, you’re able to have much more accurate and up to date analytics to use for your downstream business decisions. If you’re interested in learning more about Tamr or any of our customer case studies, please visit our website at Tamr.com and reach out to someone from the team or team. We would love to speak with you. Thank you so much for your time and have a great rest of your time.