Trigger Campaigns
Becoming Realtime
🎤 Context
ActionIQ is a Customer Data Platform that helps companies connect and unify customer data, empower flexible segmentation and orchestrate marketing campaigns across all digital and offline channels.
It is a complex system with multiple product layers used by different user groups. For such a product, bringing a brand-new feature to the current system will require thoughts on potential changes on existing parts of the platform besides building the new flow itself. Thus, evaluating the scope, picking the right consistency to keep and develop new patterns that could scale demand a lot of efforts and are design problems themselves.
One example is the Trigger campaign project.
My role and teamwork
The project involved early research, building out a quick demo and an end-to-end product cycle that involved Design, Engineering, Product, and the Account engagement team.
🔥 What’s Trigger campaign, and why now
In this fast-paced marketing world, missing a couple of hours could mean missing the conversion opportunity all together.
While ActionIQ is great at utilizing historical data (usually 24~48 hrs behind) to build campaigns with personalized content, “realtime” has always been a heated topic in numerous sales meetings and client syncs.
Realtime-ness is a crucial missing piece in our mission to help brands interact with their customers in a personalized and timely way.
💪 Goals & Challenges
As an enterprise start-up that is quickly expanding our customer base and navigating through the rather noisy market, we have 2 major goals for this initiative:
Enablement: prospects & clients
Define a feasible and scalable project scope that could check the box for sales as well as progressively deliver customer values.
Challenge: Make (sometimes bold) trade-offs on what to change on existing product and what to include for the new flow.
Unique strength: personalization in realtime world
We don’t want to (and can’t be) another version of transaction email business.
Challenge: Add the value of personalization in the rather generic trigger campaign world while working with time and tech limitations.
💻 A peak to the Minimal Valuable Solution
🐥 Start from the concept validation
Before committing to this initiative, we did a 2-week Special Project Sprint.
Sprint Goal: Build something tangible to show our prospects and customers for better concept validation
Timeline: 2 weeks
Involved teams: Design, Product, Pre-sales and Engineering
The demo was well-received among potential prospects and existing clients.
🏭 Scoping: Make it to product
While it only took 2 weeks to put this into demo, there were a lot more to consider when thinking about productization.
Scoping framework
In order to have a high-level picture of the full scope first, I did 2 evaluations collaboratively with Product, Engineering, Account Engagement and Sales teams.
Step 1: What’s the impact on existing product
Two decisions made at this phase significantly reduced the design & implementation work without sacrificing customer value and prepared the key workflow in a good shape to scale to the ideal state.
Step 2: What’s in the new flow
For this step, I put together the new flow by evaluating 5 major elements: Create, Change management, Validate, Activate and Monitor.
On top of that, I also considered Permission, Organization, Metadata and overall UI consistency to ensure the experience of this new object is consistent with other objects of the platform.
Two decisions made at this phase are good examples to show how we make trade-offs between consistency and constraints.
Things not to keep…
We decide to build the validation flow for later because 1) there are good workarounds to achieve the same goal both in & outside the platform 2) it is tremendous design & engineering work that might postpone the launch for months
Experiment with new patterns!
We’ve received a lot of feedback from users for having to create new campaigns to change small details because of campaigns are immutable. This time we decided to take the chance to make Trigger campaign mutable and apply this new pattern to other campaign types if it works.
💡 Unique value: Beyond enablement
After having a clear picture of what we need to do as the full scope, it’s time to to pick the right story we want to optimize for V1 so that:
It could help marketers do better personalization in the realtime context
We can deliver the value quickly and start learning from it
Think big, start small
Historical data could be helpful in various steps of the workflow. While it is tempting to build an exploratory dashboard to help marketers envision better marketing strategies, it could be valuable as an extra layer “filters” when building the realtime campaigns, which is lacking in a lot of traditional transactional email tools.
Thus, we decided to focus on adding strength in “building” phase (by adding “Eligibility audience” & Exporting Batch variables) rather than building a fancy dashboard for discovery purpose.
🎨 Optimize for monitoring experience
Visual story: make assumptions
From the marketing research, internal insights and past experience building other campaign features, I listed 3 valuable stories a marketer might be interested in:
📖 The basic: What’s happening now? What happened in the past?
📖 Can I trust the number: How we get to the final qualified number?
📖 Identify the patterns: Is there any interesting pattern that could inform my future campaign strategy?
I put together a few tangible ideas for each story so that we could test ourselves and with our customers.
Validate our assumptions
With 5 external interviews & multiple internal review sessions, we found out that the most relatable and valuable stories on monitoring phases are:
🙌Primary story: Show me the numbers
🙌Secondary story: How we get the number
More Explorations: layout, information architecture and visuals
It became much more clear what the ideal visualization story looks like after the sessions, I then made more explorations to see how could it could tell both stories in a hierarchical, accurate and smooth way.
😎 Outcome: MVP Walkthrough
Scenario
“Ryan is a retail marketer who wants to launch an abandoned cart trigger campaign for fashion shoes over 150 dollars to convert potential buyers.”
First, analysts and ActionIQ field team need to make data available in the platform and prepare them in a format ready for business users to use
Then, marketers setup the Trigger campaign definition including picking the source stream(Abandoned Cart), adding additional event filters (product category, cart_price), utilizing historical user data for more personalization (emailable) and re-qualification period (every 30 days). They also add one or more activation channels (e.g responsys for emails) through which they want to send the marketing communication.
Finally, marketers set the campaign to live and start monitoring. By looking at the campaign dashboard, they can get a quick grasp of how many people are qualified on the a certain time-range (the basic: what is happening), how we get to that number (build trust: clarify the process)and how many people get activated in specific channels (the output: success & failures).
Impact
Business: Making Trigger campaign part of our regular product line differentiates ActionIQ from other Customer Data Platforms (CDP)during sales pitches. Being able to execute campaigns in a “personalized” and “timely” way is a huge benefit for our prospects as well as existing customers. It’s been written in contracts and move the up-sell process forward.
Product & Design debt: Besides its business value, we also took it as a chance to deal with design and product debt. The product consistency cheat sheet was created during the process and guided us to identify the broken parts in our old system and experiment new patterns (e.g. allow edits/changes after live).
Collaboration: Last but not least, we all learnt how to identify market signal, take baby steps and scope out a big scale project as a team. In the fast-paced startup world, shipping is highly valued, but “shipping the right story” is even more important. The scoping process of this project itself is valuable practice to set a foundation for future projects in terms of how we collaborate, deal with ambiguity and make bold trade-offs with cross-functional teams.
🙌 What works
Break a large project into smaller milestones: quicker feedback loop, less “it’s going to be endless” feelings
Keep a decision doc: less “we’ve discussed it before but I don’t remember why we made that decision”
Focused review sessions: Define your design principle, name your design concept
🤔 If I have more time
Process: Look at more realtime system (outside marketing world) for inspirations
Process: Prioritize style-guide related tasks early on
Last updated
