Content design and data: a love story
How a content audit led to a 20% reduction in marketing spending, a +12% of CTA conversion rate, and ended the “what is content design?” debate.
Whether through analysis of metrics, A/B testing, or conversion rates, content designers can extract insights from data, make data-informed decisions thanks to a clear rationale, and easier influence stakeholders.
How to get there:
- Benchmark: check what others do, and A/B test it
- Backlog pain points: trust your user research qualitative data
- Backlog pain points: improve CTA clarity
- Data: tailor UX to user behavior
- Fresh eyes: “What did you expect?”
- Fresh eyes: if it’s confusing to you, it’s confusing to users too
First things first: what’s Casavo?
Casavo is the next-generation European platform that matches home sellers' needs with homebuyers’ desires through an integrated and personalized experience.
Our mission is to change the way people sell, live, and buy homes in Europe. To achieve this ambitious goal, we are redesigning the home selling and buying experience, piece after piece.
1/5: Backlog pain points
As a content designer, starting work in a new field is a nightmare, as all specialized words and terms are just gibberish. Still, you are supposed to know them all by hand because “hey, you work with words, don’t you?”. It means you have to learn. Fast.
Despite not understanding “Conditional binding offer”, “Fair market value”, and “Return before interest” (and yes, people use acronyms, so I am just being nice here by spelling them out), from day 1 people started coming at me to give me feedback about flows, glossary, and products.
I knew I had to document it all, despite not fully understanding what they were talking about. I would have gotten it out of my hat during the content audit.
I created a Backlog pain points page in my Notion (any fan, here?), including a reporter (in case I had to deep dive on that point), and copy/paste Slack messages link, Gmail, and all. Centralizing all their feedback made me feel I could get it out of my head, and work on that later.
Once you get the organization better, you will be finally able to detangle the mess by adding a project, a tribe, and so on.
PS: you don’t have to treat all feedback. Stock it somewhere just in case, but remember you are in charge: define priorities, and decide which feedback you don’t want to treat (biased feedback, lack of data, and the one and only “gut feeling”).
2/5: Preparing the battlefield
I know, you can’t wait to start. You want to get your hands dirty and start messing around. But remember — you know nothing (yet).
Before jumping to a Figma file or Excel Gsheet, take your time to look around both internally and externally.
- Get to know your company. I arrived at the company after a colossal rebranding, and I asked the marketing team to share with me the new brand guidelines — it’s like a 101 manual of me of the company
- Get to know your peers. I did too many 1:1 to count. 15' per person, with the whole product team, and the whole marketing team. Slack is good, but coffee is better. And your Backlog of pain points will grow exponentially!
- Decide on which project to focus. That was an easy choice on my side: I’m part of a tribe (Sellers), and their most impactful experience is the valuation flow. Over 190k people reached out to Casavo to evaluate their homes.
- Test the flow yourself. Many times. The valuation flow ends up with a valuation for 100% of users, but there are 7 different junctions to arrive at it. Test all scenarios, stress cases, and all errors. Make sure you have access to the staging to avoid Seller Advisor Experts (the people who call all sellers after the valuation flow) hating you.
- Ask for competitors. Do you remember the dusty Google Slides presentation that your manager randomly sent to you during your onboarding? Well, it’s your Holy grail now. Do a first general benchmark of the flow to get a general feeling, length, tone of voice, and questions asked,… A screen recording about the happy path is enough, for now, you will have time to study these flows in depth in part 3.
3/5: Getting your hands dirty — on Figma, mainly
Here is what my Figma file looked like.
It was divided into 3 sections:
- as is: copy-paste of existing production screens, so it will be easier to compare the pre-post version
- to be: updated screens according to your audit insight
- benchmark: a step-by-step benchmark to keep inspiration high during the whole process
Tip: if the flow has a lot of screens and you plan to work only on some of them, add a colorful rectangle with low opacity to let people see at a glance what you’re working on.
Benchmark: check what others do, and A/B test it
Don’t reinvent the wheel. Other people before you already worked on the same flow. It’s out there — go check it for inspiration.
Let’s take the CTA to start the valuation: as it’s the main entrance to the flow, it’s key to make sure the wording is the right one.
Competitors’ choices can help you figure out which one could be better: “Get my free offer” (Opendoor), “Start” (Offerpard), “Get a free agent valuation” (Zoopla), “Next” (Redfin), “I want to sell” (Loft), “Get your offer” (Clickalia), “Get an offer instantly” (Tiko), “Sell your home” (Rive), “Get my estimation” (Zefir), “Get an offer” (Homeloop), “Estimate” (Meilleursagents). Yes, you saw right: no two are the same!
As it’s no joke to decide on this CTA wording and our benchmark did allow us to fairly clear cut, we decided to do some A/B tests and put in the loop the brand marketing manager and the performance marketing manager. I’ll let you know which one wins!
2. Backlog pain points: trust your user research qualitative data
In the valuation flow, the number of rooms or bathrooms will determine the final valuation of the house. When looking around, some competitors had a definition of what they meant with those terms, to make sure users are feeding the algorithm with the right information.
When asking the Head of Product if they have had a similar issue, his answer was clear: during interviews, people asked “what do you mean by room?”. But what about users doing the flow on their own, without a Casavo employee to help them figure these terms out?
At first, I suggested creating a tooltip, but given the number of people asking for this info, I opted for a subtitle. But there was still one question to be answered: what do we mean by “room”? I then decided to contact our advisor Antonella to figure it out. It took 30 minutes of discussion, several forums, and dictionaries, but we created the first draft.
It’s still a work in progress because it’s way too long, but that’s where we are heading.
3. Backlog pain points: improve CTA clarity
Being someone very selective when it comes to inviting people to my place, I was very dubitative when I saw that the CTA following the house valuation was “show us your house”. The user just gave us a ton of information to let us evaluate their house, but it’s too soon to step a foot in the door.
Moreover, the CTA is not entirely telling the truth — at this step, we don’t need the user to physically invite us over, but we ask them to take some photos of their house (or eventually, to video call us). Much less invasive, literally.
We switched from “Show us your house” to a more general “Complete your house profile”. Can’t wait to let you show the results in part 5!
5. Data: tailor UX to user behavior
In the valuation flow, we asked sellers when their house was built. However, after analyzing the data, we noticed that a lot of them were entering the decade (ex. 1980, 1990), rather than the specific year (ex. 1998). Out of 31 300 valuations, 38.93%of them (12 220) had a construction year ending with a 0.
Our algorithm does not necessarily need the specific year, so we decided to follow users’ behavior and make it easier for users to input their information by switching from “What year was your house built?” and free text, to “When was your house built?” and a free text and dropdown.
5. Fresh eyes: “What did you expect?”
You’ve filled in all the information about your home and come to a final form to enter personal information. Most of the users drop, and it’s no surprise.
When they showed me the final form, I asked what happens if (miraculously!) the user puts in his data. To my surprise, they told me that it gets to the next screen, which contains the evaluation — So, no email recap? no Casavo agent calling me? Wonderful!
Why is the UI telling me something different? We need to show, not tell, that the assessment is ready. Together with the product designer, we created a drawer letting people see that the valuation is already waiting for them. Data don’t lie: the users like it!
6. Fresh eyes: if it’s confusing to you, it’s confusing to users too
When doing the valuation flow, users are asked about the status of their house. Casavo needs to know if it’s free, inhabited by the owner, or rented.
The “short-term” and “long-term” rental options did not include a timeframe, and users cannot guess what we mean by those terms.
Fun fact: when I started asking what we considered “short term” and “long term”, answers were far from unanimous. After asking to Acquisition Operational Excellence Lead and legal department, I added a timeframe that varies according to the country legislation (Casavo is available in France, Italy, Portugal, and Spain) and specific country-related needs.
4/5: Great things are never done by one
As you might have guessed, a content design audit is not a project you can do on tip-toes. In this case, the valuation flow is the main entrance to all Casavo’s sellers, so you have to involve a lot of stakeholders to make sure your changes are worth it.
Here is how I proceed:
- I communicated to the Head of Product I wanted to work on the valuation flow audit
- Once I did 80% of the audit, I tagged the Head of Product and the Product Manager to ask for async feedback for copy-related changes
- Thanks to their feedback, I put other people in the loop (ex. legal, seller advisors, BI analysts…)
- I wrote in the #sellers-design channel on Slack I was working on the audit. I pushed them to give feedback about copy in async and planned a meeting to discuss all changes that needed design rework.
5. During the meeting with Product designers, I prioritized feedback about design rework, and at the end of the meeting, we had a clear Go/No go decision. We also included those proposals in the current sprint tasks (for example, the team was already working on the revamp of the construction year screen).
For copy-only changes, I had to implement a new process for existing keys update (in 2023 there will be an article about content design processes at Casavo, hold your horses!) thanks to the collaboration with our Localization manager and translators. It’s an adaption of the current process to create new keys, but we had to implement some variables (inclusion or not of product designers, Jira tickets outside of the current sprints, impact analysis on the conversion rate, glossary updates,…).
For copy and design changes, some of them have been included in the current sprint tasks (for example, the team was already working on the revamp of the construction year screen). The remaining ones have been put in a backlog.
Getting your seat, and enjoying the show
Once it’s done, there’s nothing left to do but wait. Even if not all tickets have been treated yet, I can already show some positive outcomes of this project.
Let’s talk numbers
The CTA update from “show us your home” to “Complete your house profile” (check change rationale here) made the weekly visit CTA conversion rate spike from 15% to 27%.
The change of UI of the form (see change rationale here) made a 20% reduction in marketing spending. For the same amount of people we push to complete the flow via a performance marketing campaign, 20% more complete the flow. So we can save a good chunk of the marketing budget.
Let’s talk people
From Marketing to Legal, from Data to Sales ops, this audit let me meet different people. Despite doing as many 1:1 as possible, I strongly think that the best way to understand other roles is to work with them on real projects.
Let’s talk content design
As the first and only content designer at Casavo, I had for sure to do a “What’s content design?” presentation. But I believe “show, don’t tell” is a way better approach.
Even if some changes were related to content design principles following a strong rationale, I decided to work with my peers on cross-domain topics to let them see that “it’s not just words”. And it hit home (well, at least it seems so).
Conclusions
This article summarizes my first content design audit project at Casavo.
After a general introduction to onboarding-related topics (Backlog pain points, Preparing the battlefield), I started working on the audit (Getting your hands dirty — on Figma, mainly), and then I shared my insights and thoughts with the team to prioritize changes (Great things are never done by one).
At the end of the project, it was time to check the results, write this article to do a proper retro of it all, and just relax (Getting your seat, and enjoying the show).
PS: when writing this article down, I wanted to be extra honest and also showcase our fails and work-in-progress decisions. Remember: “one take” is a myth.
Thanks to the whole tribe Sellers, the design team, and all the stakeholders who helped me break this project down. ❤️
Have some feedback on this article? Great!
Let me know in the comments, or send me a message on LinkedIn ✨
