In this podcast, Blue Margin co-founder and CEO Brick Thompson hosts CIO John Manzanares. John is a seasoned IT leader and a board member for PE-backed field services, construction, transportation, and logistics companies, as well as the Society for Information Management. He served as Executive VP and CIO of ITS Logistics, a PE-owned $900 million revenue transportation and logistics firm. Prior to that, John was VP and CIO of CoolSys, a PE-backed $500 million HVAC and refrigeration services company. John also led as CIO of Daylight Transport, a $200 million transportation provider. With a master’s in computer science from USC, John brings over 30 years of corporate IT, private equity, and industrials experience.
“First and foremost, as a CIO, you need to be a business leader, and technology is the tool you bring to the table to help solve business problems. If you’re not creating business value with a project, then why are you doing it? If you can’t understand what the business priorities are and speak in business terms, understanding the value of what the customers and shareholders need, you’re missing an opportunity to improve and drive technology to solve business problems.” -CIO John Manzanares
Principal Interview Themes:
- How a PE-Backed Midmarket Commercial Services Company Monetized Their Data
- How Data Lakes are Key to the Buy-and-Build Strategy
- How Microsoft’s 365 Copilot Will Likely Change Data Reporting
- How to Wrangle ChatGPT in the Wild West Days of AI
- How to Minimize BI Project Delays
How a PE-Backed Midmarket Commercial Services Company Monetized Their Data
Within a buy-and-build strategy, PE partners should plan how add-on acquisitions’ data will be integrated. John advises, “You need to think ahead. You don’t start building right away. You have to think about this from a business perspective. What are we doing? Where are we heading? Build a solution around the business.”
While CIO of CoolSys, a PE-backed midmarket HVAC refrigeration company, John created an IT roadmap that helped facilitate the company’s successful exit. Within 27 months, the company acquired 14 other businesses, posing integration challenges and opportunities.
To address the challenge, CoolSys partnered with Blue Margin to build a data warehouse, management reports, and executive dashboards. These helped power operational improvements and uncovered process inefficiencies, which led to savings and EBITDA improvements of hundreds of thousands of dollars each year. Download the CoolSys case study to see the details and dashboards.
How Data Lakes are Key to the Buy-and-Build Strategy
“The shift towards using the data lake with some type of model on it, where there’s a data layer to help provide consistency and understanding – really plays into that speed to value that private equity is driving towards.” – CIO John Manzanares
Data lakes are rising in popularity as the centralized data repository for PE-sponsored companies because they offer a more agile, faster, and less expensive approach compared to traditional data warehouses of the recent past. Unlike traditional data warehouses, which require defining a data schema up front and which store highly structured historical data for specific purposes, data lakes allow schema definition after data storage. They allow the storage of raw, unstructured data without predefined purposes. Although data lakes may have slower query performance compared to the traditional data warehouse, their popularity is rising due to overall speed to insights. By incorporating a semantic layer for consistency and clarity, data lakes enable rapid insights and analysis during the critical first 100-days of a holding period for PE operating partners and portfolio operating teams.
John comments, “If I’m a PE owner and we bought this platform company and we’re adding acquisitions to it with an exit strategy of 3-5 years, do we have time to wait around for the huge effort that needs to go into that [traditional] data warehouse? [Instead], what value can we get right now from looking at data in the data lake, and getting that data into the hands of business leaders to make quick decisions?”
Learn more about the benefits of creating a data lake before investing in a data warehouse – including faster speed to insight, a testing ground for KPIs, and readiness for Microsoft’s 365 Copilot AI tool – in our recent podcast: Why You Should Build a Data Lake.
How Microsoft’s 365 Copilot Will Likely Change Data Reporting
“I’m guessing that near-future tools will be able to look at data repositories, like a data lake, and enable natural language querying (much better than we’ve had). I think MS’s big investment in open AI and their commitment to putting Copilot in all of their tools may lead the charge here. And I bet we’ll see some LLM tools replace or become the scaffolding behind Q&A tools.” – Brick Thompson, CEO, Blue Margin
(Please note that Blue Margin Inc.’s views of Microsoft products are entirely our own. Although we are a Microsoft Gold Partner, we are not being compensated to promote Microsoft tools).
The newest versions of large language models (LLMs) – especially ChatGPT driven by GPT-4, will change the way end users access data and reporting and even input code. With Microsoft’s Azure OpenAI, programmers are currently previewing the Chat Completion API model that uses ChatGPT-4 to write code. (Microsoft, 2023a). With Microsoft’s 365 Copilot, LLM is integrated across the Microsoft suite so that end users can give Copilot natural language prompts such as “Let my team know how we updated the product strategy” and Copilot will generate a response based on Outlook meetings, emails, and chat threads (Microsoft, 2023b).
Similarly, CIO John Manzanares imagines that tools like Copilot will likely be able to pull data insights directly from Microsoft’s Power BI and other data tools directly into email digests, delivering executives the high-level summarized view they want while business analysts and other regular Power BI users can still log into Power BI to filter and slice more granular views of the data. And when executives have data-driven questions, they’ll likely be able to get quick, accurate answers with simple natural language queries.
PE sponsors and portfolio executives should prepare to leverage this upcoming technology by readying their data. John comments, “What’s the foundation, if you’re doing anything around AI or ML in your organization today? Data. The foundation is data.” If your companies do not already have a data lake, now is the time to build one. It’s possible Microsoft will launch Copilot for Power BI in the coming months. If you have a data lake in place, you’ll be positioned to take advantage of the new technology immediately, possibly creating competitive advantage.
How to Wrangle ChatGPT in the Wild West Days of AI
“You talked about hallucinating – that’s making up data. That’s because these large language models, they’re not necessarily building intelligence, as much as they’re building this neural network of likely probabilities of what the next sentence is, or what the next information is. And where are they getting that data? They are feeding in massive amounts of information. And you know the internet – not everything out there is 100% correct. Some of that data is flawed, has wrong information, and is socially biased.” – CIO John Manzanares
With a lack of compliance and regulation, it’s fair to say that we’re in AI’s Wild West portion of history. Brick and John discuss current happenings and how to lead organizations wisely during these ChatGPT-4 frontier days.
The duo discusses ChatGPT-4 learning and awareness, code development, privacy considerations, corporate policy, and the need for education. John comments, “I think you need to understand more about the tool to be able to effectively communicate the challenges of using it. Go in with your eyes open, understanding the risks and the challenges.” Two risks related to AI code development are security flaws and litigation, as ChatGPT-4 can write bad code and copyright code repositories.
Privacy is another significant challenge rising with ChatGPT-4. Users should assume that all entered information may be shared and refrain from disclosing proprietary information. Like tools of the past, this tool of the future will require both policy and education. Companies should develop ChatGPT-4 and AI policies and educate their employees on how to use the tool safely. John sagely reflects, “There is no evil in the technology. It’s how it’s used.”
How to Minimize Business Intelligence Project Delays
“This project is not something that you can just give to IT. You’re the business leader. We want to build a solution that meets your needs and expectations. You’re the expert here. How can I partner with you, so this is our project, not IT’s project?” – CIO John Manzanares
IT projects (including BI projects) have developed a reputation for delays. As a seasoned IT veteran, John shares five practical tips to avoid them:
- Start with the standouts. Identify the areas for biggest business impact, the areas of biggest pain, the groups eager to participate, and the functions seeking results. Start there.
- Set clear, measurable expectations for business leaders. (E.g., “I will need five hours of your time in the early stages of this project, and your attendance at a weekly steering meeting so we can ensure project progress.”) By involving business leaders throughout the process, you avoid the concerned onlooker perspective and instead gain the commitment of fully engaged executives.
- Show incremental improvements along the way. Use an agile approach that prioritizes “speed to market” and iterative progress. John comments, “Start. Start small. Build into that. Don’t go out there and try to show perfection. Show incremental results.” (Adam Coffey, Tania DiCostanzo, Tracy Hockenberry, and Andy Scott also support this approach.)
- Appoint an experienced change management leader. This individual will rally the team, write internal communications, engage stakeholders, and generate excitement for the upcoming project. John comments that you want employees to say, “Change is something I’m part of, not something that is being done to me.”
- If you don’t have an internal team, find the right external partner with the technical expertise and experience to help you define a roadmap, establish a foundation of good data, and get your project off on the right foot. Once your BI infrastructure is well established, your internal resources can maintain the system.
Connect with John
Blue Margin increases enterprise value for PE-backed, mid-market companies by building and managing their data platforms. Our strategy, proven with over 250 companies to-date, expands multiples through data transformation, as presented in our book, The Dashboard Effect. Download your copy here and get weekly insights on our podcast.
More Expert Insights Series Interviews
- John Lanier of Middle Market Methods
- Vishy Venugopalan of Weave Growth Partners
- Brian O’Neill of Designing for Analytics
Coolsys. (2023). CoolSys | Parent of Top Refrigeration and HVAC Service Companies
Microsoft (2023a, March 21). How to work with the ChatGPT and GPT-4 models (preview) - Azure OpenAI Service | Microsoft Learn
Microsoft. (2023b, March 16). Introducing Microsoft 365 Copilot | Microsoft 365 Blog
Society for Information Management. (2023). Home - Society for Information Management (simnet.org)