Pia Wakefield on Data Cleanliness
Enterprise Data and Business Intelligence (BI) Program Lead
Emerge was at the 14th IT Financial Management Week (ITFM Week) to get up to speed on the latest in the ITFM industry. We managed to grab time with Pia Wakefield from Ciena Corporation to talk about her presentation at ITFM Week, the dirty work of data cleanliness, and her experience with zero-based budgeting.
Thanks for talking with Emerge today, Pia. Introduce yourself to our readers.
I am Pia Wakefield, program lead for Ciena’s Enterprise Data and Business Intelligence (BI) program. I also double as the Technology Business Management (TBM) point for the IT Strategy and Planning team, which is my home base.
We are speaking together at ITFM Week. What brought you to this conference?
This is my third ITFM Week conference, I think. I like this conference because it's the “in-between” conference. The focus feels in between the TBM conference with its heavy tool- and tech-orientation and the IT Financial Management Association (ITFMA) which is more structured towards IT Financial Management and classic ITIL. ITFMA keeps it tool agnostic - there is no talk about tools other than in the demo sessions, where people can showcase their tools. It's meant to be very process and subject matter oriented.
In my experience, you find more tech leaders at the TBM conference has (including the C-levels with powerful stories) and more finance leaders at the ITFMA conference. ITFMA Week seems to have a more even mix. It is a little bit more liberal when it comes to showcasing and talking about the products that may be supporting the processes. It's smallish, so it's still intimate, sports a good mix of sessions and people seems to speak freely about the applications or the support they're getting from the apps without it being a pure product commercial.
Which topics have been of the most interest?
IT transformation and how to fund innovation is a hot topic. Finding money to fuel innovation is something I think we all aim for. Even if we’re in the starting blocks, lower on the maturity scale and mostly dealing with cost transparency and implementing good fundamentals for planning currently, the end goal is to shift more fund towards innovation. Understand what we spend, what drives that cost, who our consumers are, then sniff out what leverage we have for optimization and apply those savings to fuel innovation.
You are giving a presentation here at ITFM Week, what topic are you speaking about?
The disservice you give yourself if you’re waiting for perfect data before starting a TBM or ITFM journey. In our discussion group earlier today, we started talking about whether you need to have a full-fledged CMDB (Configuration management database) before you think about application total cost of ownership. We had a healthy debate about whether you can (and should) start with what you have. Just beginning the data collection and using it in conversation may spur teams and data owners to augment and improve it. Data, however imperfect, gets your program rolling.
Everyone starts with bad or imperfect data. Instead of waiting for perfection, apply five easy steps to improve data quality over time – define your target, get a baseline, identify gaps, build a roadmap, and execute. Rinse and repeat.
At the start of the journey, help educate your data owners about what that means. Sometimes teams don’t even know or recognize that they are the data owners. The data is not their focus; how it plays in their processes may be. If they don't know that they are a data owner and that that comes with specific responsibilities, you can't fault them for not executing on those responsibilities.
IT, when consuming data from around the company, whether through TBM or otherwise, may consume data slightly differently than the team that owns that data. Things that are not a problem for them could be problematic for us (or other consumers honestly). We need to show them why we, together, need to make that better over time.
Is it an issue of data quantity or cleanliness?
It’s not a difficulty of getting the data in the first place—it’s when you get the data. Sometimes people are reluctant to share because they don't feel it's good data. Finance usually doesn't have an issue with sharing because of quality, after all the data is good enough for them to run the books, so it's useful data. That data may not be as good for you as it is for them because again, there could be holes and gaps in that data that doesn't impact their ability to roll things up and provide reporting, but could make it hard for us in IT to understand how to model or allocate the cost towards the things that are going to consume that costs.
It can be data sets that have invalid data in some fields. Some fields may hold a TBD value because the system used for input didn’t allow “no input,” but the cost isn't known at the time of input. Assigning a TBD may not impair that part of the process but may have a downstream effect on you.
Or it can be that someone couldn’t wait for someone in IT to make a system enhancement, and decided to use one field for two purposes, which poses a challenge when you are trying to parse that data for a different purpose.
It could be any quality issues: timeliness of feeds, the accuracy of data, validity of the data, data uniqueness, or the consistency of the information you need.
What’s the number one initiative you are working on right now at Ciena?
I'm part of an enterprise program to improve data governance. Data can accelerate and boost the quality of process and decision making. We want everybody at Ciena to get access to the data they need to do their jobs in the best possible way, and use data to stay on the forefront, understand our business and the competitive landscape.
We are dealing with all the corporate data across the enterprise. Formalizing master data management and precise data quality controls, baking it into the whole process of how we operate and to make the transaction more seamless. We want to set ourselves up to a higher level of automation should we choose to, or faster-traversing and speed of data through any system, with much better analytics and reliability. TBM is almost like a micro-universe of what we're trying to do with the whole company. TBM is making sure IT can be run as a business, providing IT leaders with data that can be aggregated and analyzed, distilled, and viewed to provide insights to cost levers, utilization, and consumption.
So, what kind of steps are you taking towards that now? Do you have a phased approach?
We are rounding out the first year of a multi-year approach to make data more self-service. We have a program team mapping out the roadmap to identify business initiatives that would be helped by doing data in a different or improved way.
We're trying to piggyback on initiatives to fix, improve, or implement processes and use that as the catalyst to change things on a larger scale.
We're not taking the right directive approach of saying, "Thou shalt or shalt not do these things, and it starts now" because that's not culturally what works best in our environment. We've been partnering with the business on initiatives to make sure that we get the data pieces right. As people are redefining business processes, we're also redefining the data flows.
Are you looking at it from a shared services concept? Chargeback even?
Enterprise data may be the ultimate shared service, but it transcends IT, and we are not looking at it from a perspective of charging back. The data governance program has to do with providing the company with the data we need to be more efficient. When the concept of a vendor or customer is talked about, everybody's talking about the same thing, and the same piece of data can flow from A to Z and always mean the same thing.
It's just visibility.
It's truly about improving the data quality for all the data. A customer address could be used to route communication, deliveries, or to drive the correct taxation on an invoice. An address is not random and should never be a TBD. A delivery address may be different from a postal address—that’s perfectly fine. We need to know what’s what and be consistent.
The data and analytics program is an enterprise-wide program that wouldn’t normally be sourced from IT. But because the business didn't necessarily have resources with the right level expertise, we are running it from IT with strong sponsorship from the business. The business owns most of the data, but they don't necessarily always know that.
The enterprise data governance program is separate from our TBM initiative, but it will help with TBM success. I'm trying to evangelize that, getting our IT and finance teams up to speed and helping us run IT as a business using TBM processes and tools.
How are you and your organization thinking about financial management in the context of projects and agile planning and analysis?
We are in the middle of an agile transformation, but we are not tying it as much to the financial side as we are trying to track direct metrics for speed of delivery.
Our budgeting planning is still very bound to a yearly cycle. We are not doing rolling forecasting yet, that's part of our roadmap to have better insight in IT—even though those numbers aren't necessarily going to be rolling back out to finance more than twice a year. There isn't a huge focus on the financial aspects of being agile yet. We're still crawling towards walking!
So, many people at ITFM week here have been talking about the struggles with zero-based budgeting and starting instead with hybrid budgeting from a baseline. What has been your experience?
We are using a hybrid approach. We're aiming to use zero-based budgeting for anything discretionary and baseline what’s inevitable: depreciation, multi-year contracts, fixed labor. Those things automatically roll over. Everything else we are trying to make people think about building from scratch. We expect this to be much easier going forward because we are using Apptio IT Financial Management Foundation.
Do you see foresee eventually getting to zero-based budgeting for everything or is a hybrid state an excellent place to stay in?
Unless you are OpEx only, I don't see how.
Again, look at contracts and depreciation. Without making any decisions at all, X-million dollars of next year's budget is already spent - locked into contracts or existing depreciation schedules. Depreciation either needs to be written off for a one-time hit, or it's going to be rolling until your assets mature. These are a couple of categories that naturally lend themselves to move forward as part of the core of the run rate budget, right? Everything else should be assessed and viewed. And even those run rate items should be reassessed, but it needs to be done in the context of the planning you are doing. What are we going to do when this contract runs out? What are we replacing, renewing, or consolidating? Then it becomes not just a yearly planning function; it becomes more an ongoing planning function.
Absolutely. That’s our time. Thanks, Pia.