Podcast S2 Ep. 14 – David Masri

Season 2, Ep 14 of my podcast with David Masri, Founder of Gluon Digital and author of Developing Data Migrations and Integrations with Salesforce 🤓

The data nerd in me totally geeked out at this conversation!

Data was my first ❤ early on in my career.
As a DBA, I understood databases, and knew how to get what I needed.

SOOOOO much neater than dealing with hooman beins. 🙄
People, emotions, politics – they were so messy and so difficult to understand.

In time, I learned how to understand people and stepped into the world of #ProjectManagement. It also meant that I would get further from manipulating data and configuring systems… so this was a rare treat!

David’s journey into the #Salesforce ecosystem is a fascinating work, starting in the #ERP world, diving deep into complex integration and data migrations with various systems and platforms.

We talk about security and various aspects of #DataManagement and #DataMigration as he goes through his book.

I hope you enjoy this as much as I did recording it!
One for the data nerds out there!

You can buy David’s book here: https://www.amazon.com/Developing-Data-Migrations-Integrations-Salesforce/dp/1484242084/

Spotify link here: https://open.spotify.com/episode/5LkF4OKtpOoaB1kvqyfzpl?si=yvibvpSxSA2XORv97P_I2w

#OnThePeiroll

https://youtu.be/mv10phsrtW0

Transcript

Pei Mun Lim 0:05
Hello, and welcome to another episode of OnThePeiroll, my podcast where I talk about all the things that interests me, such as leadership, project management, consulting, delivering quality products when great people. And this podcast this particular episode is really fun because I get to indulge my data nerd side. I say to David Masri, he runs glue on. And he’s done a whole load of data migration data transformation projects, and he’s even written a book. So I was very cheeky. And I got him to divulge a lot of secrets. Well, actually the points within his book on how organizations should think about data migration, and why it’s so crucial. It is the one thing one workstream, I believe that makes or breaks a project, because why would you have a CRM system? You want information and business insight that can help you make strategic decisions? Now will if the information is not correct, not accurate, not timely, not complete, then what’s the point of the system? So that’s why I really, I think I could have gone on a lot longer on this particular episode, to be honest, but there’s a lot of value in there, especially for anyone looking to embark on a project because data migration is a really key part. I hope you will enjoy this as much as I did making it.

Hello, David, welcome to my podcast OnThePeiroll. It is a pleasure to have you today. So how are you feeling right now?

David Masri 1:58
I’m good. Thank you for having me. My pleasure.

Pei Mun Lim 2:02
I am I’ve been, you know, excited to talk to you for a while for a few reasons. I’m not going to get into it now. But there are few things about the journey that you’ve taken to where you are today that I would like to dig into, mainly because I’m a data nerd and one of the tutors few things that crossover that wanted to ask. So to start this off, it would be lovely, if you can just tell us the story of how you got to where you are today. And the journey that you’ve taken. And you know, feel free to share any anything, any stories that you want to tell.

David Masri 2:39
Yeah, so I’ve been in tech my whole life, and I’m not a spring chicken anymore. So it’s a bit of a long story. I’ll try to keep it not too short. Not too long. I mean, so I’ve always enjoyed technology actually took a coding class. The first time I used a computer actually was in sixth grade I think was an IBM PS two and let’s get started off with a cartridge coding in GE basic at the time. So I kind of enjoyed it right off the right off the bat and I continued fiddling, you know with computers eventually got a computer at home. My dad bought one you know through high school and into college in high school. We’re now talking I started high school 9090 So when high school 1994 So in high school I was gearing up thinking I wanted to go into business and finance and then I went to business school post high school and I liked enjoyed computer so as minoring in information systems and actually majoring in finance because that’s what I thought the world was going and in the late 90s All of that changed and I didn’t particularly enjoy the finance side of things particularly the heavy math facts aspect. Not that computer systems has much less math but still so I switched my major and my minor so I went from majoring in Finance minoring in nation systems to majoring in information systems and then minoring in finance and then yeah, they I was working part time I wanted to I wanted to you know get some exposure get some job experiences working part time through college. You know, various jobs very low paying, but it did you know computer work, either setting up QuickBooks or networks or website Access databases, this kind of stuff. So I really I really enjoyed it. I took a database class in, in college, by far my favorite information. Class I keep in mind that time, computer languages were much less richer than they are now. Right hacky way. Java was version 1.1. And I cannot stand that thing at a time. So I really liked databases. I actually still have my college textbook, I think somewhere. It’s the one college textbook I didn’t throw out. It’s upstairs. I gave it to my son. Three. It’s not here. Yeah. So yeah. And so I decided at that point I wanted to go into into data work, I wanted to be DBA at the time. So anyway, I finished school, started working for a media company doing lots of number crunching, mostly mostly campaign ads, crunching those numbers to performance in them. You started doing some internet work, a banner ads, links, text links, that kind of stuff, crunching those numbers. I then went on to work for Prudential Financial for a couple of years, as a DBA. By the time I was at Prudential Financial, we had the whole.com blow up. So the space changed quite a bit. And then 911 happened and things got quite a bit of work in, you know, in the economy space, putting aside the politics side of that, and in the tragedy, of course. So I found myself struggling to find a job, it wasn’t easier. So fairly Junior. At the time, I ended up joining my uncle, who was writing custom ERP systems, he had a small professional services shop. So I ended up joining him. And I stayed with him for quite a few years, five or six years, doing custom ERP systems with him learned a tremendous amount from him. That was my first introduction to professional services. And the consulting side of things really, really tuned my communication skills. My understanding of business, ERP is a great place to really, really learn business. Because you get the accounting side, you get the operation side. And back in the day, the thought process was everything in ERP, right? So yeah, so I really learned a tremendous about about professional services about, you know, Client Relations and about computer systems. And they’re also obviously heavy, heavy data systems.

Tired of integrations there, so we integrated with everything, everything you can imagine, websites started popping up, all over the place, eBay was becoming big, or was big already Yahoo stores with a thing think they’re gone now. But we’re heavily integrating with things like that. Amazon came around and started integrating the Amazon for some of our more retail oriented clients. Lots of EDI work, which is still around EDI for wholesalers integrations with warehouses with all of the various shipping companies, right, just a ton of that kind of kind of integration work. I ended up leaving my uncle for various reasons. You know, we still get along great and have a great relationship with Him that kids nothing personal there. It’s just the direction and technology who’s using was aging. It was a custom keep in mind use custom ERP, which is not is not an easy space, right? You’re not talking about implementing, you know, SAP or something, which perhaps we could have pivoted, but that’s a whole other story. I then spent a few years in the CRM space, pretty Salesforce, so is doing sales logics, lots of migrations off act. That was all C sharp coding for the most part at the time, sales logics was land based, and that moved to web based and then eventually to AWS to the cloud. So again, lots of great learning experiences over there. And then I did that for about five, six years. Then it went to to Weight Watchers. And I did their dynamic CRM implementation and call center and munch special projects, been two years of Weight Watchers and then after that, I entered the Salesforce space. I joined a company called redkite. And then there are there are a really small boutique, CRM company Salesforce shop did heavy heavily financial services, space. And then after a series of acquisitions and a call a couple of other small moves. So I was at red tape for a few years red tape got acquired by liquid hub which would have got acquired by Capgemini. I did a I did like six months at Silver Line. Great people over there loved working with them as well. And then I went to a company called GSD which event slowly became creative. They’re doing real well, they joined plaintiff as director level. And then in. In the last year, year and a half or so ago, I went off and kind of decided to start my own firm doing specifically data work for Salesforce s sighs. So to think a little bit of a step back, when I joined, when I joined redkite, I joined as a project manager, I had no Salesforce skills, all of my tech skills were heavily on the Microsoft stack. And even Honestly, when I joined wastewater, there was as a project manager, too. But somehow, when you have strong technology skills, you always get sucked back into coding. So that has happened to me throughout my career, honestly, even when I left my uncle, he was coding heavily in FoxPro. That was one of the reasons to have to leave because FoxPro was no longer supported by Microsoft. And we had to essentially to rewrite the whole platform and start from scratch, which wasn’t something I wanted to do. So I ended up joining weightwatchers, that the project manager, then he gets sucked back into the tech side of things. Then I went to redkite, again, joined the project manager, I got back sucked into the tech side of things. And at some point, it just decided to kind of, to kind of embrace it. So I had really, really strong data skills, like you said early on, he was always interested in it a bit of BI work heavy, heavy integrations. And that’s really, really, really where I was able to help in a lot of these Salesforce partners, where they had really great Salesforce skills, but not necessarily great data skills. And what happened was, I was working on a really, really, really big migration for a firm that will not allow you to mention their name. That’s a whole other, that’s a whole other thing. But that’s just how they are this particular firm. And that was not the greatest data migrations, I was also building all the integrations for this company to it was a really big project. And I was here full time. So I had a dual role I was heading the data track as well as doing the project management work. Right. So the integration work was going fairly well. But the data migration was just an absolute nightmare. One of the reasons for it was a homegrown CRM system, that was just an absolute mess in the head, no one in the company who knew the system or the back end. And it was to the point that no one wanted to touch it. Right? Well, we still had to get all the data out of it. So and as you can imagine, the contracts were written that they own getting the data out of it, and they owned the data structures, and they couldn’t deliver on that. So we ended up having to take that over. And then there was this whole crazy manual data loader process that I like, I just couldn’t absolutely believe that this was the normal way to do these kind of migrations, it just seemed so foreign to me. And then I would do research online, like how everybody does it. Right? That’s how like, if you Google, how do I do data migration in Salesforce, it’s dataloader CSVs fee lookups. A very heavily manual process. So I, towards the end of that, it started using a lot of the real strong integration tools that we’re using to leverage and fix up the data migration, and kind of smooth that out. And then when, when that when we got past that project, and it had now some time to free up a, a basically took over and reworked entirely top to bottom how we did our data migrations at red pipe. And then when I went to various other firms, I took those processes with me. Eventually, I ended up running a couple of data depart divisions within the firms, really training people and putting the best practices in use. In in 2019, I published the book, developing data migrations and integration with Salesforce, which has done is done really well and very happy with it got a ton of five star reviews, I think that 40 of them, almost averaging four and a half stars. So yeah, I mean, that’s kind of how I got where I am now. So then September of 2020, September 2020, I started Bhuwan digital, which, again, is is rather than go off and start, you know, a another data department at another firm that was struggling with data, I figured, let me build a company that can focus just on working data integrations and data migrations. And then we partner with SI is to kind of take this off their plate, let them focus on the Salesforce work, and we’ll focus on the day to work and it’s been going really really, really well though Yeah. So the only thing that I’m absolutely set on with go on is we do not compete with our clients who We do not do end client work, all of our, all of our clients are either Salesforce s eyes or or ISPs, and they own Salesforce work beyond the day to work for them. Yeah, that’s a mouthful. 20 years, five minutes.

Pei Mun Lim 15:16
Thank you know, that was quite comprehensive. We come from a similar space. So Microsoft is well over here. And I was also a DBA for a bit. And I did a lot of infrastructure stuff. So network servers, change servers, those sort of things before moving into CRM. So one of the things that really struck me was Your Weight Watchers stent, in 2000, in seven, so I don’t know where Weight Watchers began, whether it was the UK or in America. But in 2007, I ran a discovery for them. So I was aware about how manual and paper heavy the processes were especially around for, I don’t know what they call it, you know, the different locations who are running meetings and things like that. So that was really, really interesting to me that you, you know, you kind of picked it up and implemented it in later.

David Masri 16:25
So when did I start? Wait was not six, six years? I think it started Weight Watchers. 2012.

Pei Mun Lim 16:31
Yes. So no, six years after I did the discovery,

Unknown Speaker 16:35
I see. Oh,

Pei Mun Lim 16:38
for them, so I don’t know what Yeah, so

David Masri 16:39
I’m not quite sure where you’re saying they were on paper. So a lot, a lot, a lot, a lot happening. Weight Watchers is really, really interesting story. It still is Monday, go back and key get caught up with them, the people there. But most of the people I worked with also left the company, there’s one or two people still there. So the Weight Watchers, Weight Watchers is what they call them, the stores, right, they still have their vocations. Though they had that that was their core of their business for a long, long time, I think was actually started in the 30s by someone here in the US, and then they expanded internationally. So the stores were heavily paper, we actually got them off our know, they rolled out a tablet. And they also rolled out Weight Watchers online to compete because they will get they were absolutely getting killed in the market by mobile apps. Right, just a whole market was changing up nobody likes to do anything in person nowadays. It’s funny. So they so they were they went through a whole kind of digital transformation, they started with watches online and had to put up a mobile app. And then they had another project to to digitize the stores. And then the only work I’ve done with that was we will routing, you started to use the CRM to track messages between the Weight Watchers representative in the stores and their I guess, their constituents, the people who are going to the meetings. So that was but otherwise it wasn’t really part of, of the CRM system.

Pei Mun Lim 18:28
So it sounds like maybe the different locations was quite involved in how they manage the IT system. Because I’m pretty sure I was at the UK headquarters. And what would happen was that I can’t remember if they call it stores then. But yeah, each of these locations would mark down the weights of the members, I guess, and sheets of paper. And those were shipped into the call centers who would have to key them in. And the project that I was there to do the discovery on was to turn all of those into using Microsoft Dynamics. So that was where I came from at that time, and have those available and to replace the people ones. So but then I finished a discovery and they picked up a piece of work and contracted people to do the implementation. So it’s interesting to see

David Masri 19:24
it’s a completely different CRM system. It’s not the same one. I think the store system was called champ at least here in the US, and then we integrated with champ so we were heavily marketing and call center. And then we so we would use some of that data for things like you know, hey, you haven’t been to a meeting. When you come back, that would be like a store message. Then there would be like if someone was flatlining in their progress, right that they become a In a risk to leave, right, or if they haven’t gone to a visit. So we would bring that data in a very summarized way in for to be used on email marketing for retention.

Pei Mun Lim 20:13
Really interesting to see how, you know, different places evolve their own set of requirements for one at that time. Like I said, JSON six, you know, mobile things mobile app was not, I don’t think I had a mobile phone at the time just wasn’t a thing. So the first priority was to reduce the overhead on the paper management.

David Masri 20:34
So they were disrupted three or four times. And then they found new markets also. So that what they found was, what they found was by going digital, also, with weight watchers.com. And the mobile app, they were able to have much more success in the men’s market. They were much more likely to want to use an app as opposed to go to a meeting. So yeah, it’s yeah, it’s really interesting stuff. And then they had to struggle to differentiate between theirs, which the paid platform and something like Fitbit, which is free, right? Yeah. So yeah, it’s an interesting thing.

Pei Mun Lim 21:19
The how business changed, and so fast. Yes. Okay. So back to your I like data. can use I’m just going to your book. Can you tell me when the idea to write it came to you? And how did you? How long did it take? And how did you manage that alongside your day to day work?

David Masri 21:47
So I was sorry, I was actually working at Silver Line. So blind had what they call OKRs objective key object, the next key results, and they focus you see, they push it up, you know, put some personal goals on there, too. I’ve always wanted to write a book. So it was like a long time, I wouldn’t say call it a dream, but it’s just something to kind of do more like a bucket list. type thing. So. So yeah, I put it down as an OPR. With my manager at the time was his name was Eric Nelson, really, really supportive of it. I, I already had training programs in place, right, because I’ve been doing this from a training perspective, because again, I built the team at Red Kai, prior to this, and then when I went to join Silver Line, I was doing a couple of training courses on you know, on how to how to how to how to do a proper data migration, at least how I defined a proper data migration. I don’t want to knock anybody or say they weren’t doing things properly. And then we moved on to integration patterns. And I thought I had all this course material when it was clear that I already had the contracts for a book at that time, once I kind of put it together. So what it did was it put together a proposal, I sent it over to a press, they had it on their site. And yeah, I mean, they jumped on it. The reason I went I went to a press was as I weigh basically one on Amazon, it’s a boy who’s selling Amazon books, sorry, who’s selling Salesforce books. And as basically a press and packet are the only two competence packet is let’s say a little bit more of a self publishing type style, although I don’t know if they would classify themselves like that. A press, as many people know, is is a, a known, you know, a known technology brand company highly respected for a long time. So for me, it was a no brainer that you’re going to try them first. And yeah, they got back to me almost immediately. They were very interested. It took me about 10 months to write it in the proposal I had. I had like I had the whole outline, right, all the chapters, and I honestly didn’t even deviate from that too much. So yeah, I mean, it was 10 months on Sundays a little bit during the week if I had a slow week at work. And yes, basically it can only put in the work. Again, I had all the content for the most part. It just getting it down on paper.

Pei Mun Lim 24:39
Talk to me about your emotional journey during the time. How did you feel from like beginning to end?

David Masri 24:49
Senator, interesting question. I mean, I didn’t get frustrated. Writing it at all. I knew exactly exactly. What I wanted and I had a provision for this whole book, I had a, like, one of the things, when I wrote this book, that I wanted to make sure that I’m not doing it as almost a guiding principle is, I’m not taking stuff that’s readily available online, and just rehashing it, you know, it’s all brand new content. It’s not stuff you’re gonna find anywhere else. It’s, it’s a patterns and practices book, right? That’s a subtitle. So again, developing data migrations and integrations with Salesforce patterns and best practices. So there is some code in there, but not a tremendous amount. But and it’s it’s tool agnostic. So it’s not like it’s pushing you towards your particular tool. I do discuss tools, obviously, because that’s important. But again, it’s not guiding you towards, towards a particular tool. And again, it’s all it’s all patterns and best practices, right? This is how you design your code diff, which flow, these are the core attributes you need to be you need to be aiming for, and then how to structure it. I’m really, really happy with, with how it came out, maybe probably should have spent a little bit more time on the grammar checking and the editing and all that you’re never happy with. You always always managed to find the typo somewhere after and but other but otherwise, otherwise. Yeah, I mean, I wrote that book in order. It’s not like it went back and, and modify chapters. The only time that I did do that was, I think it was up to chapter four or five, when I wanted to change the style of the book from lecturing where I’m talking about me and you to, to more of a weak, like, let’s do this, or we need to do this. Right, just to be a little bit more inclusive, and, and more friendly sounding nature. That was the only time I had to go back and like just the end, but even then was very systematically go back and change the the style of writing. People at least one person, my tech, my tech editor, Jared, his name is real great guy. Yeah, like he thought that I used, like a, like a dictated it and had a computer. Just, you know, just record what I was saying when I wrote it. Because it’s a very conversational style book. I wanted to keep it relatively light in terms of for easy fun reading. Yeah.

Pei Mun Lim 27:37
So in terms, do you want to share just something if you were to summarize that book, to someone who you know, wants to buy. So I, for example, I’ve not watched that, but I was going to cheekily ask you for an autographed copy in how I might get it. But to see someone who’s thinking about moving into Salesforce, and I want to tell them, you want to get Dave’s book, it’s got, you know, can you just summarize that?

David Masri 28:11
Yeah, I’ll take you through how it’s structured. It starts off, leave it or not, chapter one is relational databases. So I cover the foundational stuff that if you’re coming from the Salesforce side of things, you’re probably you may not know, right? What’s an index? What’s, you know, what’s the basic join the philosophy behind the relational database? Why it was designed all of that kind of stuff? Right? So that’s Chapter One is just understanding relational databases for people who don’t come from a database background, right? Chapter Two is for is for people who either know Salesforce, or don’t know Salesforce, and it says Salesforce tells you it’s a relational database, but it’s really not. And this is how Salesforce is structured. The differences between Salesforce and a relational database, and then we go into how Salesforce is kind of architectured. And then. And then why, like, Why did Salesforce make the decision to move away from standard relational rules? When they did, right? So that’s chapter two. And the reason that’s like an important chapter is because your most people they come from Salesforce, they, they think of it as a relational database. It’s really not. It’s close, but it’s not. And it’s great to compare it to a relational database, because that’s the closest thing to it. And the other thing is, when you’re building code with Salesforce, and you want to understand why am I getting locking errors, why am I you know, why is it designed this way to understand that the architecture and the theory behind the architecture is very, very instrumental in that and that’s what a very According to right and that’s why I included that chapter chapter three, overview the Salesforce API’s, right and how to use them. And then some basic introduction to creating fields. And we do a little bit of data loading with the like to the exercise to do some basic data loading with the Data Loader. Just to get you understanding there. We also cover other ETL tools and for the cons of those chapter four, I start getting into best practices. And then chapter five best practices, I’m sorry, the core attributes, I think Chapter Four is the core attributes. I’m going to open the book because my memory is not as good as I think it is. Or is it was a copy was a copy. Okay, so Chapter Four goes into the six attributes of a good data migration. These you’ll probably see my content on my blog, Salesforce data, blog.com have a ton of writing to supplement the books even though after the book finished, I wanted to continue to expand on a lot of topics that are somewhat out of scope for the book, but we touched on it in the book and then really wanted to build on it, that chapter five is the attribute of good integration, right? And then six is best. Chapter Six is best practices. So go through about 40 best practices and relate each of the best practices to the attributes that I outlined. The six attributes essentially is well planned, automated, controlled for first of all repairable and testable. For a good data migration. Integration, the difference is, repairable is not enough, it has to be self repairable, meaning because it’s going to continually run, it needs to fix itself if something goes wrong, not rely on on a human interaction. But chapter six, again, best practices, tie them back to the attributes.

Chapter Seven is putting it all together. So we do a sample migration, start to finish, chapter eight error handling and performance tuning, really interesting chapter, I have a great blog article on performance tuning, that was expanded from this chapter, or summarized down from this chapter. It’s something that people really don’t understand the inner workings of the Salesforce data engine. So we kind of cover that, in this chapter, chapter nine, we’re now moving towards integration into synchronization patterns. How do I synchronize data between two systems? I think we go through like six or seven patterns of how to do that. And we talk about the pros and cons of each of them. Right. And again, this is not going to no code in this chapter. There’s like visual flows and whatnot. But the point is, is to understand how to think about data synchronization, and what are the challenges? And how do you avoid conflicts and all of all of that kind of stuff. Chapter 10 is other integration patterns. So these are things like how to deal with record merging, how do we deal with ETL based roll ups? How do I archive data out? Right, so either not data synchronizations other chapter 11 real time integrations as well as UI integration. So not just data, maybe I want to integrate between various wise user interfaces. That’s the core of the book for the most part, everything else is essentially extras. So chapter 12, is I have a whole library of reusable code, which I share that either are SQL functions for the most part that do common transformations. For example, I need to parse out the domain of an email address and put that into the web fields, right. So I need a, you know, convert camel case because my draw date is all uppercase. Like lots of these kinda kind of code snippets that can be reused in a sequel environment. Chapter 13 is FAQs. Just go through a bunch of commonly asked questions. And then I have three appendix is Appendix A. It talks about data cleansing and deduplication. Again, not necessarily the core of the book, but in a really, really important topic. And it goes through an algorithm for detecting duplicates chapter. Appendix B is references that just like reference cards, if you wanted to pin them up or print them out, and that kind of stuff, just to get quick reference guide. And then chapter Appendix C is just referenced it in the reading further reading. So just every single, every single footnote in the book is summarized there. And every single data source or source that I’ve used, so if I said, this is how the Salesforce engine work. They pointed to the Salesforce doc that that came from. So yeah, I think this is on my blog, too. I ended up pulling out to the appendix and making it a blog post as well. Um, that’s essentially the book See, like I said, it’s, it’s very structured. And it’s not stuff that you’ll find generally, in other places.

Pei Mun Lim 35:10
That is incredibly comprehensive. I wish I had something like that when I was doing, you know, my data migrations. And I didn’t do very much integration, because that means a lot on coding and things like that. So, in the Microsoft side, I just stayed with SQL databases and things like that. So that is pretty comprehensive. So in terms of just trying to think in the UK, we’ve got this big thing around data privacy, and when we will, you, there’s GDPR. And therefore, data migrations used to be a big pain. So what is it like in you know, in America in terms of handling personally identifiable data? How do you handle it as part of the projects that you do?

David Masri 36:00
So this is an interesting question. It’s funny, because I’ve often do a lot of data security work, you know, for my clients, and previously, when I was at various companies, just because you happen to have the expertise. So over here in the US, and I’m not a lawyer, so I apologize if I get any of this wrong. It’s generally only only health data that’s really, really super private, right. And then you also have other data that’s really strict, because it’s a real security issue already dealing with financial firms. And maybe there’s regulations on, you know, big trades that might happen, that kind of stuff. But the PII is not necessarily a huge issue, you might have bank account information, which again, can be a big disaster if it’s if it’s compromised. So essentially, my philosophy is, and keep in mind, and coming from the professional services side of things, is I have a set of security requirements, the bare minimum that I’m going to meet, our clients have the st have a set of requirements, and we’re going to take the higher of the two. Right, so. So if the client has super strict requirements, it’s going to be no problem, you stand up the infrastructure, you give us access, we work on your, on your hardware, right. You know, our goal here is not to become, you know, a data platform, but we’re storing data for people and, and really taking ownership of it. You know, that being said, you know, we do have some of our own infrastructure, which again, I had secured myself, it tends to be for clients that have not as strong needs, or wants in terms of their security, or ours is just stronger than theirs, because they’re, you know, relatively small shop. And then very important, have a, make sure you’re doing your due diligence when it comes to data destruction, right. Don’t leave stuff out there. When you don’t need it. That’s probably the number one thing you can do is write don’t keep data that you don’t need.

Pei Mun Lim 38:19
And so, you know, in your very long history, you will have done quite a number of projects. I’m keen to hear about your favorite one, do you have a favorite project with a client that you think might you that wasn’t really the project.

David Masri 38:36
So keep in mind that my clients now glue on our, our Salesforce Si, so it really, they’re not my clients, they’re their clients or projects that I’m working on. So I’m not necessarily comfortable talking about the specifics. I will say from an integration side, generally, my favorite projects tend to involve really big data warehouses that need to be stood up with an analytics component, and then some kind of integration with Salesforce. And yeah, those are those are, they tend, they tend to be, you know, very enjoyable and really bring a tremendous amount of value to our clients. So the other one that I think those are the the categories that they find the most interesting and the most challenging. In terms of data migrations, in a lot of way, the data migrations are harder than integrations and a lot a lot of ways and people don’t realize that. The biggest the biggest reason is it’s not uncommon to get a data migration that I need to migrate 25 objects 30 objects, that is incredibly uncommon for an integration and integration is one or two objects, maybe four or five and that’s pushing it Right. And then at the end of the day, it’s got to be wholly automated. And it’s still got to work. And I gotta do my error logging, right. And it’s got to run in a short period of time. So, you know, data migrations, again, a lot of times a lot more, a lot more complex, they do get a little bit more on the tedious side of things, particularly on the mapping and the data cleansing and all of that stuff. So yeah, I think I definitely enjoy the data integration. Work a bit, a bit more, particularly for, I guess, I would say somewhat, a little bit less, technically sophisticated, not no, not nothing. But like, you want someone you know, a company that’s, you know, somewhere in the middle, they understand technology, you know how to use it, you know, they can take advantage of it, but they’re not there yet. And then you can really, really get them, you know, to, you know, to be a company that’s really, you know, data driven. Yeah.

Pei Mun Lim 41:01
So, what’s was, was there a specific point in time? Or was that trigger that made you think that you might prefer, you know, starting I’m blue on at that time, what made you make that jump it’s the combination

David Masri 41:28
or the combination of things, essentially, essentially, I was already thinking of moving on from my, my current employment, for whatever reason. And then I had a couple of former colleagues of mine, different companies, basically saying, hey, come work for us more for us. And I ended up just basically taking them both on so I had a had a clear path by so rather than go working for one of them, each side also worked for both of them, and do my own thing. And then I always had this idea of, you know, maybe I should try to build a, you know, a data migration division, that can help us size. And this was the ideal time to do that, because I now have the freedom to without competing with anyone. But that’s kind of how it happened. Call it a victim of circumstance, you know, how it is opportunity. Not, it’s not like, it’s not like it’s about different not like, it’s something that I was planning for years and years and years, I probably spent my whole life saying I don’t want to work for myself, because of, you know, the amount of work and not necessarily the payoff, and yeah, it’s it’s long hours, a lot of a lot of stress as you can imagine. So yeah, it’s not necessarily all that it’s cracked up to be working for yourself. And again, it’s not something that necessarily wanted to do, I kind of just ended up doing it. Anyhow. I’m not one of those proponents that say everyone should go out and work for themselves. You know, if you want to, if you like to find there’s nothing wrong with it, obviously. But there’s there’s definitely, definitely downsides to it as well. I’m also not exactly a materialistic person, I’m probably the least materialistic person you’d ever meet. And the money is absolutely not a driver. For me once you know once I’m paying my bills or pay my mortgage, I’m not worried about things right. It’s like there’s just you know, it’s not gonna buy happiness at all. Like nothing. So yeah,

Pei Mun Lim 43:54
so tell me about how you’ve positioned yourself to serve as the Si Si so what what what does it look like they go out and get a project and they just outsource the whole day to make pay Peace to you.

David Masri 44:10
So that one of the things I wanted to do when I started we won is not constantly be chasing sales because I’m not a sales person by nature. So wanted to develop partnerships where I can get reoccurring business you know, from the same firms minutes also great way because you build you build friendships that way to you working with these people regularly, which is nice. So they they generally feel like colleagues more than clients for the most part. The other thing I like to do is I really do like training people. So though, yeah, you know, I struggled at the beginning finding, finding the right, the right employees. I recently partnered with a company down in Uruguay called The EFG. And they’ve been absolutely great, right? So what’s really nice is they have an office that that people that can go to, you know, he’s able to, you know, he knows the, you know, the local community and either local culture. Right. So so it’s, it’s great, you know, for hiring people tend to think that, you know, having an office to go to is not is not important. But I found that it really, really impacted my ability to hire, particularly, particularly offshore. Right. So, so yeah, so So I partnered with this company, Vfg. Right, they helped me find people, they hire people that and they also have a deep bench. And I’m then able to train them up, oversee quality control, make sure everything’s running good quality, I can do most of the client interaction myself, as needed, and it’s been working out great. I’m super thrilled with how things are going. I think that was your question, I’m not even sure anymore.

Pei Mun Lim 46:07
I just wanted to understand where you how you are framing your offering. So

David Masri 46:16
that’s, that’s essentially it, I partner with an SI. They have a project that needs a data migration or an integration. Or they asked they, you know, asked me, you know, you got capacity for this? And I say, you know, either yes or no, generally it’s yes. And then we, you know, we take that on. It’s essentially that we act as an extension of their team.

Pei Mun Lim 46:41
So you do you do the discovery bit as well in the beginning? Or do they do that, and then just posting user stories and poems to you?

David Masri 46:51
Know, we tend to be involved in the discovery as well. For the data bits, we’re not necessarily doing hardcore Business Discovery, right? Again, our goal is to stay out of their space, right? Let them do what they do well, themselves, and then we come in there and help. So for example, for data migration, depending on the client, some client really wants to do the mapping himself and handle that. Or we can do the mapping, but we’re still gonna need the assistance of the Salesforce VA or consultant because they on Salesforce bill, we’ll do the data analysis of the source system, right? Because again, they generally won’t have the skills for that. And then we go back and forth with our questions, just like he said, no different if you were working with an internal data migration specialist, right? It just, it just no different than having that role internally, we act as an extension of it. So

Pei Mun Lim 47:50
right, so let’s say hypothetically, you’ve got this hypothetical, because I did have this client who needed to so there’s a few different things. One was they needed to move off their retail platform that was creaking. And the license was expiring. And a whole host of other things that they needed integrating and migrating as well, challenge was one of the legacy systems. The vendor, the vendor had gone bankrupt, they kept going through, you know, goodwill, in lots of prayers, and the administrator left. So from a partner point of view, that was a huge risk, because they kind of required some reverse engineering to try and figure out the data model and things like that. So in terms if that was the scenario, you would be the partner who would come in and have a look at the legacy system, am I right to look at how

David Masri 48:58
Yes, so so. So here’s, here’s at a high level, the process or how I would approach your exact scenario. Step one, I want to assist them demo user interface this right? We’ll go through the through the whole thing, and we’ll take notes on okay, this is what’s being done. This is what the system does. This is what’s being captured the data that’s being captured, at least at an object level, understand the one to one relationships, the one to many relationships. And then we get access to the back end. However, that is of some API. If it’s a sequel, backup, if it’s an oracle system, whatever it is, we get access to the data. And we we devise a plan to extract the data, all the data, okay. Again, if it’s a sequel backup that might just be able to take a backup and restore if it’s API’s, maybe I gotta write code that extracts all the data enough and loaded into a local system repository. That repository serves two purposes. One, it’s going to be a source for for migration. And again, that extraction code is automated. Right? So it can always extract it again later, it’s going to be the sources in for migration. And once you shut down the set the the live system, that save backup, in case they need that data later, for whatever reason, you know, maybe that’s patched on migration, because the defect was found later. Maybe it’s, you know, they decide that they didn’t migrate something that they need. Also, maybe it’s they need evidence for a dispute, I don’t, right. So it could be for whatever. So you have that, then what we do is we do data analysis. So I now know what data I’m looking for. We then go and we find that in the in the data repository, and we usually do that by querying the metadata, right. And for the most part, it’s not that difficult if you know what you’re doing. There are occasions where some systems just have like, generic names custom field, one custom field to custom two or three, custom field text one, custom field text two. And usually, when that’s the case, what you do is they’ll have the metadata somewhere that’s driving how that’s displayed on the screen. So it’s really cracking that code. So we’ll spend, we’ll spend the time to understand the legacy schema. And if needed, the metadata that’s driving that if it’s a system that allows for heavy customization, right. And then we assemble a mapping doc from there. And then we’ll we’ll move for move from there, we often validate the mapping doc by the UI, right? So I have a saying, Okay, this is what I believe is the address field. This is what I believe is the ID, this is what I believe it’s the name, I will go into the UI, you search for it. And then you can literally validate the data match at the data match at the data matches. You’ve had an instance like sometimes you just can’t find a piece of data, we’ve literally written code that searches the entire database, every single field for a particular string, and then you’ll find it, right. So again, you know the things if you come from, you know, from a strong data background, they don’t sound so outrageous. But when you’re coming from the Salesforce world leader, not skills that you find within the Salesforce world. And again, this is one of the reasons we started to go on is to really bring these kinds of skills to the table. Anyhow, at that point, our data migration is wholly automated, right? We have the source data, we have a transformation layer that transforms it, transforms it to the Salesforce object model. And then we have data that, again, code that loads it loaded loaders, no manual steps, and it just literally, kind of push button from there from the extraction transformation to the load, then we do our validation for the client. And then if the client finds defects, we don’t fix the data, we fix the code that’s transforming the data and loading it, and then we can rerun it. And that’s how we know that when we’re ready to go live, this thing is going to be spot on, right? Because there’s no, there’s no human processes that can you know, slip a finger or make a mistake. You know, one thing that computers do do better than then people must be consistent. And that’s essentially what we’re relying on. And I think that’s really where a lot of manual data migrations fail is one day, they, they consistency in the human process. Right? So they’ll so what happens is you have you have

to have happens all the time, right? I have a heavy data migration project, they sign a Salesforce ba loaded data in and like, here’s an Excel sheet of their accounts within a Chelsea with a contact with an Excel sheet of their opportunities. They look at it no problem, right? spend a couple hours they bang it in. Right, ready to go live, they do the same thing they validated. Right? And they go ahead, they do it again. They do that again, the next project, right maybe to have activities this time also, no problem. They go they bang it and this person then becomes a guy who knows how to do data migrations, wonderful. Maybe he writes a blog article, you know about it, and then they’re put on a $2 million project where they have to migrate 30 objects, they’re giving databases they’re not giving CSVs they’re dealing with millions of records that no computer and a laptop is going to do with Excel is going to crash. They can’t do B lookups and they end up with like a, a document like this long e nine pages long, Do this, do the transformation, do that transformation, do the transformation. Right? load this file, get the success files merge the six files, right. And it’s it’s literally that long, they finish the process to do the data migration takes three days. So there’s no going live on a weekend, right? They tell the client, you need to be offline for two and a half weeks, they get a defect, they’re not going to spend four days running the whole thing that they just patched the data, then they update their notes. And then they’ve never done a full end to end process, right? And then they wonder why the thing is failing. And then the the person who’s supposed to do this go live quits, because they’re terrified of it. So they find the new job before they have to go live. And then somebody has to inherit this. Right? Yeah, it’s it’s, it’s an absolute mess of a process. But it happens all the time. All the time, you hear the story over and over and over again. And that’s really like right to the the pain points you want to you want to address. And it’s not this bas fault, right? I’m not blaming this BA in any way. They were not trained to work with data. And they should not have been assigned it to begin with the problem is that they had a couple of easy tools and a couple of easy migrations that gave them confidence. And then now it’s like, you know, it’s a completely different, you know, different ballgame. So, yeah,

Pei Mun Lim 56:30
absolutely. I think the process that you described is just magic. Absolutely magic. And not enough partners and companies and projects, view data migration, in that way, is basically a patch job. You know, like, let’s look at this thing. And how do we do this? Just my greatest thing, how about everything else, all the interconnected bits, just not thought about? So thank you for sharing that that’s really, really useful. And I kind of wish that it was more. So you said, you said earlier on that you enjoy training is training one of your offerings. So if let’s say a partner in the UK, is listening to this, and think, oh, you know, what, we just need to level up our data migration capabilities, is training something that you’ve arrived?

David Masri 57:24
It is, it is, it’s not a core offering, because requires so much of my personal time. But I have absolutely done this, like I said multiple times in the past. And again, it depends on the team that you’re that you’re shoring up. The other part of it is it takes close to a year to train someone to be good at data migrations, and they depends on their varying levels of skill. So for someone who has, who has decent data skills, meaning they come from a data background, they know how to work with databases, they know how to work with SQL, then you gotta teach them how to deal with Salesforce API’s. And they gotta learn the Salesforce object model and all of that kind of stuff, right. And then, you know, to have someone who knows them, both Salesforce and SQL that tend to not really need training much. But you don’t find those people too often. Very few and far between. So it’s either usually taking a Salesforce training, and then you got to train them on core databases and SQL and that takes a while, or you’re getting on with solid SQL skills, and you got to train them on Salesforce. And then both of those people need to be trained on data migration, best practices, and the patterns. Again, it’s all outlined in my book, so you can always self train.

Pei Mun Lim 58:56
Sounds like you’re providing a very niche, you know, kind of, you know, the cross section of all the skills that’s required to do this properly, especially complex projects. So thank you for sharing that. They are very mindful of your time. One fun questions before we wrap up. You said earlier on that you couldn’t find the book because you had just given it to your son, how old is your son? And what’s he doing with your college book?

David Masri 59:25
So it’s interesting. So he’s currently he just finished high school. He’s 19. It’s going to be 90 and actually next week. So he took a gap year spend it in Israel. And then he’ll start he’ll start Brooklyn college next year, and of course, scholarship honors. So real proud of him. I don’t know if he wants to go into technology. I was starting my company and he wanted to get get some part time work. So I said, What on databases you know, then I’ll give you a job. Is there something along those lines um, But it was before he went off to school, but it’s still in his room is always it’s one of the nice things about it about SQL, I really hadn’t changed that much in 20 years. Nor has relational databases, at least, the fundamentals.

Pei Mun Lim 1:00:18
Thank you so much, Dave, for that, I really appreciate you taking the time to record this. For me, it has been really fun, because, as I mentioned, you know, I am also a data nerd, and to hearing your journey and your story, especially how you’re using your skills and your experience to tackle a very important gap GAAP. So I’ve seen a lot of data migration issues, problems. And I did have one question. So before we, before we go, you were talking about how data migration, you know, might take a while. So both have backgrounds coming from Microsoft, where, you know, the service on premise, databases, it yeah, um, migration doesn’t take as long as it does now, where everything’s on the cloud. And there’s a, you know, it’s slow, really,

David Masri 1:01:11
the bots built force, again, I have a whole chapter on performance tuning. Salesforce is really good. With loading massive volumes of data. The issues that you face are the custom code, which are essentially database triggers. Right? That kick off as you update records. So if you if you were to load, let’s say, you were to load a flat object that has no lookups, no roll ups, no formulas, nothing on it, you can load millions upon millions of records in an hour, for sure, using the bulk data and not get a single, right, the issue is the locking, and then the database triggers that kick on. And there’s various ways to approach that. I talked about how to avoid locking errors pretty extensively. And that’s essentially the biggest thing. So performance tuning in really, really, really super summary, avoid locking, they discuss a couple of that in the book and in my blog. Second key, don’t push updates you don’t need, that’s obvious, right, in terms of performance gains. And then thirdly, if you have automations, that are slowing things down, you, you want to know if you can turn them off, sometimes you can’t. The common advice online is oh, just turn everything off. That’s not necessarily true. Because you might need whatever that automation does, you have to understand the automation. And then to if you understand the automation, might be able to turn it off and then update whatever that automation was doing in your ETL code. So for examples, roll ups is a great example. roll ups are going to be slow in Salesforce. So for example, if I’m pushing in, you know, hundreds of records, or hundreds of 1000s of records, let’s say, and each one is triggering a roll up. That can be very slow. But if I push all the updates, turn off the roll up, let’s say right now, I know you can’t necessarily turn off roll ups, but you turn off the roll up. And then you have the ETL calculate roll up and then update those records once. Right. So there, there are ways to go around that. I mean, in this case, again, you can’t necessarily turn off Salesforce roll ups, but he told me about a cool baseball, or like roll up helper helper type thing. Yeah. So again, it’s the issue is not going over the internet. That’s just you know, internet latency. The issue is primarily, you know, database triggers, trigger type logic, right workflows, process builders, that kind of stuff. And then locking in the big thing can Salesforce does a lot of this stuff. Beneath beneath the hood, even when you have no even when you have no code. So for example, updating a record with a master detail is gonna be slower than updating a record. Right. And it’s again, because there is trigger codec just

Pei Mun Lim 1:04:16
amazing. Thank you very much. And I also like the fact that you’ve written a book. So anyone, anytime someone asked me a question, it’s in chapter three, that’s in chapter nine. So but thank you for taking the time to even summarize a lot of that I totally relate. And I have enjoyed this conversation very much. Once again, I thank you for your time, Dave. And perhaps at one point, that might be a part two, there are things that I want to talk about it more, but let’s leave it for now. I really appreciate your time.

David Masri 1:04:48
If you want to do a follow up, we’ll be more than happy to and we should figure out how to get you an autographed copy.

Pei Mun Lim 1:04:54
Oh, that would be amazing. Thank you, Dave. Thank you.