Designing for an Information-rich Dashboard and Concurrent Workflow
A few months ago, our team in PebblePost set out to rebuild the Programmatic Direct Mail platform from the ground up. This included the backend infrastructure, big data analytical mailing engine and the digital campaign manager user interface, to fulfill the need for business growth. The first version of our platform was so successful, that we outgrew the capabilities and capacities of the entire platform. . Therefore, we took this opportunity to rethink what we need to build to better support the user’s needs and business’s goals. As the lead and solo designer in PebblePost, I want to share the challenges we faced when designing the new version of PDMD(Programmatic Direct Mail Dashboard) and how might we overcame them.
At PebblePost, we combines the power of digital intent data with the effectiveness of in-home marketing to achieve far higher conversion rates for businesses. We invented Programmatic Direct Mail. There is a great deal of engineering and data science efforts behind the scenes to make this all happen, but here I want to share my work from the perspective of product and UX design for the new version of PDMD.
Designing the Dashboard
PDMD is not only a dashboard that merely displays data or shows charts. It is a comprehensive digital platform, by which users can organize campaign orders, set up PDM campaigns, manage and optimize campaigns, measure campaign effectiveness. And its capabilities are expanding.
The Challenges
The uniqueness of PDM is that, instead of relying on static demographic consumer information or third party user data, it captures real-time online intent signal data and translates them into manageable “strategies”, which users can leverage and reuse when setting up campaigns. These strategies are constantly being updated and optimized base on non-stop intent data feeding(the “Programmatic part”) to make sure the in-home marketing media(the “direct mail” part) sent out are sufficiently targeted and delivering the best results.
While our first minimal viable product has validated the core value proposition and product/market fit. We also identified some missing pieces that are critical for the product to succeed. The new version we are designing is not just an incremental improvement or UI renovation. Based on the grounded learnings of what value we are creating for customers, we are looking for the most appropriate and intuitive UX/UI patterns to present the underlining PDM technology, which has undergone significant overhaul.
For the new release, we are trying to address three major problems for users
How might we enhance to productivity?
With more capabilities comes more control, therefore more effort needed to set up a campaign. It is an expert system with a lot of sophisticated settings. Users need to take time to set up campaigns according to their business goals. The quicker we enable marketers to set up a sophisticated campaign and launch it, the earlier we can prove our value and the more business we can win. One behavior we observed is that users keep cloning campaign setups to save time on repeated entering. But process is tedious and error prone.
Reusable Settings as “Shared Ingredients” to pick from
The observation of user’s cloning behavior implies there are a lot of repeat use cases for setting up a campaign. For example, users want to set up another campaign, which targets the same consumer segment but in a different time frame. They would clone the whole campaign and only change the date range for it. It saves a lot of time reentering the same information, except that it is dangerous if they may forget to update the offer code accordingly. Sending promotion to consumers with expired offer codes is a big no no.
Instead of letting user go through a big chunk of settings and rely on cloning repeatedly, we break down campaign setting into reusable pieces. Users can save the configuration of who, when, what, how respectively as “ingredient” in different libraries, then they can browse and pick for the appropriate ones when setting up new campaigns without having to enter repeated information repeatedly. The “ingredients” can be shared to a different “chef” in the same “kitchen” to improve collaboration and productivity.
How might we streamline the complex workflow?
To set up a valid campaign, users need to configure the appropriate elements of When, Who, What and How for each one. These elements are interrelated and some of them depend on each other. This nonlinear process usually leads to missing configuration, conflicting settings and resource misuse
Concurrent workflow and Guided process
When envisioning this reusable setting solution, we realized that not all “ingredients” are 100% compatible with the new food, some may need minor adjustment before applying. For example, in the PDMD, a “strategy” setting defines a collection of “who” to target, which can be reused by different “flight” settings. The “strategy” itself is composed of different “segments”, which are also reusable settings. Most of the time, users can select strategies and directly apply them to flights. But sometimes, creating a flight is not a linear process. If the user wants to add a new segment to the strategy before apply it to the flight , they have to pause creating the flight, create the segment, add the segment to the strategy, go back to the flight and eventually they can apply that strategy to the flight.
To simplify the workflow, we imagine the entire workflow can be composed by several linear sub-workflows. These workflows can happen concurrently. Human behaviors are naturally multi-threaded. A chef can prepare other side dishes simultaneously, when they are working on the main dish. The key is to remember the progress of each dish so that they are not missing ingredients or over applying them.
Firstly, we need to make sure each sub-workflow’s status and context is maintained unambiguously for users so that they can keep track of progress and not get lost.
Secondly, we provide each sub-workflow a step by step guidance this is both beneficial to first-time users and experienced users. The guidance is presented as ordered steps visually but users are also free to jump around. New users can use it to proceed through a new setup more linearly. Experienced users can jump around more freely to check if anything is missing.
How might we lower users’ cognitive burden and prevent mistake?
The nature of PDM fuels the system with a lot of dynamic information. Data are constantly flowing through the system, big data processing may update the status and state of various elements or resources asynchronously. We believe in data-driven, in which more data points will help users make more informed decisions. However, we don’t want to overwhelm users. How do we effectively display information to help users to make sense of it?
Signal System: Status, Alert, Confirmation
DMD enables users to leverage the most up to date information and reusable settings to set up and manage campaigns. However, it inevitably creates dynamics inside the system. For example, a proven effective strategy is likely to shared by many flights. An update made to this strategy will cause a chain of changes to all relevant flights. We need a robust status system to present important states, an alert system to notify users promptly, and a confirmation system to force conscious actions. The goal is to provide users with the most critical information in an unobtrusive way and make users feel that they are in control.
Foundation UX Patterns
The goal of our first new release is to test a solution that can let user set up a PDM campaign.
The typical workflow of setting up a PDM can be as following:
- Build a campaign with existing reusable ingredients (Most straightforward): 1 → 2 → 3 → 4 → 5
- Build a campaign and a strategy concurrently, and apply the strategy to the campaign: 1 → 2 → 3 → 3.1→ 3.2 → 3.3→ 3.4 → 4 → 5
- So on and so forth
We want to specifically focus on
Enhance productivity, Streamline Workflow and Prevent Errors
To address the challenges above. I drafted 4 foundational UX patterns: Reusable Component Picker, Guided Process, Concurrent Workflow, and Robust Signal System.
Next Step
Coming up with the design is just the first step. And these solutions are definitely not the only ones. The good news is that we got something functional that users can start testing with and providing feedback on. We keep our iterations lean, agile and continuous. Hopefully, I can share the learnings, experiments and improvement again very soon. Stay tuned!
