FOLLOW

FOLLOW

SHARE

The Key To Building An Effective E-commerce FP&A Team

How Hotwire re-built their FP&A team

16Jan
Matt is one of our confirmed speakers for the FP&A for High-Tech Summit! Get more insights from him by joining us in San Francisco on April 18 & 19. View the event.

Today’s online e-commerce environment is ever-changing and ultra-competitive. A financial planning and analysis (FP&A) team must to be able to partner with its business leaders and meet growth objectives while managing the complexity that an online world entails. Most models that are applied to build an FP&A team address a structure, but do not prioritize its implementation. Based on my experience, I found that I needed to apply a model that helps prioritize data and reporting as the foundation so that the FP&A team will be successful on delivering its vision as the trusted e-commerce business partner it strives to be.

Case Study – Re-building the FP&A Team at Hotwire

I came to Hotwire in the spring of 2015 to manage its FP&A team: my fourth division within 4 years at Expedia Inc. This was a fantastic career opportunity; however it quickly presented itself with significant organizational and functional challenges. Organizationally:

1. The FP&A team was decimated – two team members out of five quit right after I took the job, there was no permanent team lead in my position for 6 months, and the CFO left shortly after I started

2. Business partnering was limited at best – none of the leadership team members had confidence in reported results and to them, the financial forecasting process 'was the most broken process at Hotwire'

From a functional perspective, it didn’t get much better:

1. Hotwire Finance was loosely integrated with its financial systems – Excel was the tool to manage all of the historical financial performance with large interlinked files and no backups, while only final model outputs were uploaded into Hyperion (not making use of its modeling capabilities)

2. Metrics were inconsistent with the rest of the organization – there was no comparability internally across brands making Hotwire unable to leverage the corporate infrastructure

3. Scorecards were incorrect and unclear – business drivers were not broken out for visibility, while key dynamic assumptions were incorrect and hardcoded in the BI systems

4. Intercompany revenue share management was needlessly complex

These challenges all together represented a crisis and I needed to address them in a way that ensured the FP&A team was set up for success in both the short and long-term.

Why is FP&A different in e-commerce? It’s simple: data.

Today’s e-commerce business operates in a dynamic environment that creates a proliferation of real-time data. This data provides a significant opportunity for organizations to make decisions rapidly to react to customer trends, and drive business profitability.

Analytics teams were the first to take advantage of this data to develop leading operating indicators which helped direct the business’ product management and strategy. This is what led to the 'test-and-learn' philosophy that has been the foundation for many internet platforms and helped explode the 'data science' domain. FP&A teams tried to follow suit to use this data to ensure it had a business view of profitability, but the traditional FP&A processes, systems and skills were often ill-equipped to address these new challenges: they were slow, reactive, and linked leading to lagging indicators poorly. The e-commerce FP&A team needed to change by really focusing on the most important criteria – data:

1. Data is fundamental to the FP&A function – without it, you’re lost

At a technology infrastructure level, the massive amounts of customer data flowing through its shopping path captures product, channel, conversion and purchase data. This complexity creates a huge challenge to ensure valid and relevant predictive models across these multi-dimensional attributes. If not managed appropriately, the vicious cycle of garbage data-in, garbage data-out can lead to missed opportunities, masked expectations, and can significantly impact the bottom line.

Compounding this complexity is that multiple views on the same data can create chaos – one version of the truth is necessary for management clarity and consistency, as data governance usually resides in finance to ensure this singularity. And there sometimes needs to be separate views of operational data and financial data – not being able to distinguish the two can add to this lack of clarity in decision-making.

Companies who rely on 'big data' will be challenged to provide insights if they are not properly set up for data management; they will be flying blind due to its complexity and lack of organization.

2. E-commerce businesses operate in a rapidly-changing environment where real-time decisions are critical

Having the speed to deliver is important to make timely decisions; e-commerce business environments can change rapidly due to the competitive nature of online marketplaces. Being set up for success requires being efficient in getting to the right information as quickly as possible – automating data processing and trend analysis are essential to solid decision making.

3. Revenue and marketing cost modeling are critical to scale profitably, while capital investment modeling is less critical

A major goal of an e-commerce business is to achieve scale with limited fixed costs. However to increase volumes, you need to manage the short and long-term trade-offs of the revenue from customer transactions vs. the costs of customer acquisition & processing. Infrastructure investment is important to ensure the business keeps up with current technology, but it is relatively fixed compared to the variable nature of an online business, and can also be managed in a less time-sensitive manner.

Ensuring operating profitability can even be more daunting if revenue and/or costs cannot be recognized immediately, as estimates of future profitability are more variable requiring more and more granularity in the metrics.

Due to these underlying forces of the short feedback cycle in e-commerce, FP&A needs to quickly generate insights from the transactional data, apply financial profitability metrics, ensure targets are met while resources are adequately distributed, and then adjust the forecast accordingly. At Hotwire, we had the challenge of not being able to address the e-commerce feedback loop in a timely manner due to manual data collection, incorrect scorecards, and inconsistent metrics management. It was fundamental to address these challenges as quickly as possible.

The Pyramid of E-commerce FP&A Functions

Given this perspective, I wanted to explore how I could apply these concepts to current FP&A organizational models to ensure my team was set up for success. I found that models have been developed at the London FP&A Board (more analytics-based), FTI Consulting (service delivery for COEs), and at the Association of Financial Professionals (applied key domains of self-assessment). The most relevant for e-commerce I found was the AFP model; however I did not see an order of prioritization of its self-assessment domains. I knew that I wanted to scale my team’s FP&A capabilities in data management first, given its foundational importance. This is where I created key functional FP&A areas and prioritized them into a pyramid as follows:

This diagram shows that data needs to be firmly provided on a clean and consistent basis. Once that is done, I need to ensure that the right people are in the right roles to follow strong processes, as it will help automate the analysis of business performance trends. However, without proper data, a “rock star” talent won’t be able to proactively drive insights and/or meet expectations.

Next, the team can apply proper insights from the correct reports and processes to help forecast their product volumes, revenues and costs, as well as analyze business cases to develop promotions, release new products, etc.

FP&A can then apply these analyses and forecasts to share in high-level business partnering relationships and drive the outcomes necessary for success. The pyramid guided me to focus on the foundations at Hotwire: first on data and reporting to ensure correctness, next to put the right people and processes in place, and then onto higher value-add activities such as forecasting and business partnering.

Stages of Evolution within each Pyramid Layer

Understanding where the team is capability-wise can help determine how to approach the team’s growth. Each of the layers in the pyramid has a range of competence, from novice to expert. Similarly, the CEB did some research at Newell-Rubbermaid that showed how specific core capabilities can progress from novice to expert levels. This type of leveling can be easily applied to the above pyramid model:

While building from the ground up, it’s important to note that a team does not need to be at an expert stage in a low-level layer before building competence in a higher layer. Depending on the complexities of the business, the skills within the team and the financial health of the organization, there may be a need to build functions vertically in sync. However, it is not possible to be an expert in one layer higher than another having lesser capabilities. For example, you can’t be an expert (or instead more advanced) in business partnering while being a novice in data and reporting because the fundamental insights needed for strong business partnering aren’t being generated to be shared with the business.

However, it is possible to be the opposite – an expert in a lower layer while still being a novice in a higher level, such as analytics and forecasting. Improvements can be made in forecasting because the foundations for success have already been laid. Building competence in each of the layers is non-trivial and takes time – do not assume that this can be done overnight.

Applying the Pyramid at Hotwire

At the start of my tenure at Hotwire, it was extremely challenging to have solid forecasting and business partnering discussions with the leadership team without having the necessary foundation in data and reporting. What we did was while straightening out this function, we hired the right people and ensured they were proactive to help the business – these new employees could then become familiar with the challenges and help address them while they became acclimated to the business.

Our forecasting and analytics capabilities are still limited due to the slow progress of our data management transformation, which included metrics correction, and migration of our data from Excel to Hyperion. However we have improved significantly given the clean-up we had to make by using interim systems and addressing our process gaps at the start. As such, the team is now more proactive: operating with more automated and repeatable processes due to integrated systems and better equipped to drive strategic value. Consequently, the business is now relying on the FP&A team to drive insights and strategy for investment and growth.

Despite all that we have accomplished in over a year and a half we still have a long way to go. We must continue to strive for that trusted business partner role as it requires building on the foundations of each of the successive layers of the pyramid.

Conclusion

Transforming an e-commerce FP&A team requires building in standardized functional layers on top of clean data and reporting. At Hotwire, I was able to rebuild our FP&A team by prioritizing from the ground up on effective data management and reporting up to predictive analytics and forecasting in order to provide the indispensable insights to our organization’s leadership. As 'big data' and machine learning begin to take over the internet, our FP&A team not only will be well-prepared to address this challenge, but it will provide our business a leg up on its competitors.

Comments

comments powered byDisqus
Data culture small

Read next:

Building A Culture Of Data

i