Acosta Retail Reporting

Acosta Reoorting

A company-wide single source of truth for automated and intelligent insights 


Execution and compliance analysis
Opportunity surfacing engine 

Single source of truth
From scratch data model 



Dozens of current reporting tools
Excel based current system
Different clients, different expectations
Multiple hierarchy levels 

My Role

Sprint Facilitation

retail reporting process

Initial Understanding

"The only useful data is something that helps you make an informed decision."

Unfortunately, that was not a mantra that was currently instilled into the Acosta reporting ecosystem. I collected every current report I could uncover and found there were roughly 81 individual and custom reports that I could document. Of those documented, rarely did any two reports contain the same KPI’s, nor did any have consistent branding, layout, or visualizations.

At first glance, each report seemed to be totally separate from the next (even within the same client or family) and on the whole, reports did not appear to have a consistent scheme of visualizing data.

Thankfully, the business knew there was a problem, unfortunately, they didn't know how to fix it in an intelligent way. This is where I got involved.

reporting current nda

Research & Interviews

"In the interest of what we thought was providing good service, we adopted what our clients wanted us to do instead of thinking about how we should be doing reporting."

This style of thinking lead to a disparate landscape of excel based reports which were often times too unusable, slow, or inaccurate to fully utilize. Needing a more detailed understanding, I interviewed 15 employees across all levels of hierarchy within Acosta to determine what the main sticking points were with the current system.

- - - 

"What we have today is a duct tape solution."

"Reporting now is 'rearview mirror facing' meaning we are not focusing any efforts on the future, nor providing prudent actions to anyone."

"Proving our value through reporting insights is not a cornerstone practice at Acosta; we do not emphasize reporting at any stage of client interaction."

The current system

reporting currently

The proposed system

reporting future

User Understanding

We knew that this was meant to be a holistic tool but to flesh out and test our ELP (Earliest Lovable Product) we needed to boil it down to some primary users.  

We needed to understand who would get the most value out of a tool like this, who would take the most action, whose lives would be made easier. We built out 11 total personas (6 internal and 5 external) and decided upon two primary groups who would receive the initial launch. 

Internal – Retail Business Manager1
Internal – Account Analyst1

Mantras & Ideation

"There’s no such thing as information overload, only bad design."Edward Tufte

We started the design efforts with a roundtable with stakeholders running through sketching and ideation exercises aimed to determine the true reason behind each page, section, screen, and interaction.

Along with those sketches and ideas came design mantras that would guide not only design efforts but development and marketing material.

- - - 

1. Make it easy to quickly measure the health and performance of the team.
Highlight items that have the chance to lead to a valuable impact 
Minimize items of less importance 
Help people figure out what is going on by highlighting issues, big changes, or unexpected results (outliers)

2. Make it easy to determine the opportunities that will have the most impact.
Not all data is created equal, make sure that the important information is highlighted with actionable insights … not digging for data

3. Focus on a fast, consistent, and complete experience.
Make it easy to navigate, filter, not get lost, to dive in, etc.

4. Strive for high Signal to Noise Ratio.
Don’t candy up the interface with unnecessary elements
With the smaller screen size, skip typical Excel charts & graphs that typically bad signal to noise ratios


Round One Usability Test

We met up with 5 of our original interviewees in the RCM user group and lead them through a series if tasks in an initial usability testing rough. Each task was subjectively scored using the following key:

5 = Completed with ease

4 = Completed successfully

3 = Completed with difficulty or help

2 = Did not complete, abandoned task

1 = Did not complete, thought they did


Following their test, we asked each participant to fill out a brief survey about their overall feelings towards the product.

They answered 9 questions using a Likert scale of 1 to 5, with 5 being Strongly Agreed and 1 Strongly Disagreeing.


Round One Synthesis


UX / Visual: The visual design and information architecture have had overwhelmingly positive reactions.

Efficiency: The quick-hitting insights are allowing people to glean the information they need without deep data diving.

Heat Mapping: Participants are responding extremely positively to the simple Red-Yellow-Green heat map.

Email: Email appears to be an effective way to quickly and automatically relay important statistics. Participants liked the potential reduction in clicks this gives them.

“This would make my job easier in telling the client’s story and tracking overall performance.”

- - - 


Sharing: The ability to share is currently lacking in development, so participants are defaulting to what it is they already know. (PowerPoint and Excel)

Store/Rep Level Data: Participants are pointing to objectives, and the rep completion, therefore, as a large indicator of success across the board.

Configurability: Participants do not feel they currently have enough control over what it is they are seeing and using.

Overall Health: Participants are having a hard time understanding the 'Health' analogy and the elements that make it up.

“Clients want to see the actual objective, who responded to it, and what they responded.”

Validate and Iterate

With a focus on accessibility and accuracy, the next round of ideation quickly started rolling.

Along with our accessibility fixes (colored text mainly), we understood that the design direction was strong, however, the details were weak. We needed to reexamine how we can quickly surface the most prudent information with as few clicks and scrolls as possible. We also recognize the necessity of personalized goals and thresholds to allow our participants to understand what caused a "red, yellow, green" signifier. 

- - - 

Thresholds and Goals

Observation: When presented with a data point and its relationship to a set goal, participants often seemed confused at how the goal got set, how to change it, and why it was needed.

Recommendation: Offer more affordances into the way and reason that rules and goals are set and allow for the creation of a new rule/goal.

"Overall Health"

Observation: Participants are having trouble identifying the necessity and building blocks of the “Client Health” analogy.

Recommendation: Rework the major header of the summary page to present opportunities and successes in a more obvious and useful manner.

Date Range Selector

Observation: While participants were able to change the date easily, many stated that quickly switching between date ranges to find trends would be far more helpful than a dropdown.

 Recommendation: Sharing mechanism is not currently built out. People are returning to what they already know.

Dashboard 1.0
Screen Shot 2018-09-25 at 7.53.23 AM
Screen Shot 2018-09-24 at 10.19.04 PM
Screen Shot 2018-09-25 at 8.01.44 AM
Screen Shot 2018-09-24 at 10.13.12 PM
Screen Shot 2018-09-25 at 7.59.29 AM


WIth a round of usability in the books, and with new and validated designs ready to be implemented for a more robust round of testing, we began looking into ways to pass off our designs to development.  PowerBI had been chosen as the tool of choice by the business for reporting across the board, and while it is a supremely powerful data modeling tool, it does not easily take in coded visual components. To solidify a clean handoff and to strengthen the overall relationship between design and development, I created a PowerBI Theme to correspond with the visual decisions made for the tool and created a PowerBI Master File to serve as a manifest of information about when, how, and why to use a certain PowerBI component.

json cap
PBI Style Guide V10 (1)
new style guide

As of this moment

Our development team has built out our data model, our architecture and visualizations, and is preparing for our initial launch.

With a subset of prepared users (people we have been utilizing as SMEs through discovery, design, and development) our team will be launching an ELP (Earliest Lovable Product) of this tool within the coming weeks. Check back for updates! 

Mana - Empowered Job HuntingBranding, Reseach, UI, UX

Acosta Retail ReportingInternal Tool, UI, UX, PowerBI

TransamericaUI, UX, Web Design

Rapid Response Supply ChainInternal Tool, UI, UX

Acosta E-Commerce PortalInternal Tool, UI, UX

Sam's Demo AppInternal Tool, UI, UX

AR Try OnsIdeation, Research, UI, UX

The Social RouteBranding, Internal Tools, UI, UX, Web Design

Kaya FinancialBranding, Research, UI, UX

The Repair CoachUI, UX, Web Design

Acosta UI ToolkitInternal Tool, Research, UI, UX

Boulder College CounselingUI, UX, Web Design,