Opportunity

Photographers want but don’t have time to write blogs. Focal did not have blog capability and wanted to be more competitive with their offers.

Solution

Focal created an AI-assisted blogging tool to help photographers increase their blogging output and attracting new subscription sign ups.

Impact

Self serve user activation rate increased from 10% to 70%.

1.6 blogs made for every total user.

Timeline

June-October 2023

Tools

Figma, VScode

Team

Heidi - Product Designer

Kimi - Product Manager

Codefire - Outsource Dev Team

PART ONE

PART ONE

The
Problem Space

The Problem Space

The Problem Space

What is Focal?

Focal is a business solution for photographers who want a one stop shop for website, booking and gallery set up and management. They offer subscription tiers to meet the needs of different business maturities. As a local start up, they pride themselves in exceptional customer service.

The opportunity cost of blogging

The Customer Success team conducted video calls to do client discovery and identified that the main goal of having a website was to increase leads, rather than just displaying their work.

Improving website SEO could help gain leads, and the easiest way to do so is to blog! However, most photographers felt the effort put into blogging was not worth a marginally higher search result ranking.

Photographers would rather reallocate the effort to blogging to managing existing bookings.

The business goal

Focal wanted to increase new user subscriptions, and increase use of the admin section (a place to manage bookings and galleries all in one) because it was Focal's competitive advantage. By working with the Success team to add the hotly requested blogging feature, and incorporating trending technologies like AI, we could expose the rest of the admin tools aka the competitive advantage.

Creating a new blog feature addresses the photographer need for higher website search ranking. Incorporating AI would help capture their interest and give them the efficiency they need.

note: diagram is simplification - there were other ways to meet business goals and customer needs like improving the onboarding process or providing more text templates in photographer booking pages, but AI-blogging met the most criteria.

PART TWO

The
Approach

The Approach

Breaking down the information architecture

The main path was with prompts. These would help inspire ideas for blog topics, and requested input depending on the prompt so that at the core, the photographers ideas would shine through.

The extra AI path - dubbed Instablogs - was to accommodate the needs of our photographer demographic - those that are more visual-focused, less writing savvy, and have a great Instagram feed.

This case study will focus on going in-depth on the prompts path because it addresses the most basic blogging needs.

Imagine this user journey:

Jimmy is a busy freelance photographer and ends up neglecting his website SEO. He doesn’t want to miss opportunities to gain potential clients through online search, so he tries to increase SEO by putting up blogs on his website.

As I explored, I found the bottleneck tends to be around idea generation and writing, so below I flesh out how the flow would progress.

Let's mock stuff up!

One of the hard constraints given to me by the CEO was to use existing design ideas/modals from the previous designer. This was because these designs were already partially built by the devs, and we wanted to stay resourceful.

For this case study I'm going to focus on these most basic flows:

  1. Choose your type of blog [Custom Prompt]

  2. Decide on the prompt for your first draft

Choose your type of blog

At first, the prompt-based blogs hadn't been decided as a path so there were only two options to design for. After gathering UI inspo, and I decided to flesh out V1B because the copy with CTA would help the user decide, and having no divider allowed the blog options to feel equal, with one not better than the other.

At first, the prompt-based blogs hadn't been decided as a path so there were only two options to design for. After gathering UI inspo, and I decided to flesh out V1B because the copy with CTA would help the user decide, and having no divider allowed the blog options to feel equal, with one not better than the other.

Eventually the CEO requested I add an extra option for prompts, so I repurposed the name "Custom" for it instead. Meanwhile the from scratch option became "New".

Decide on the prompt of your first draft

While we provided prompts, we wanted to give the user the chance to create their own prompt in case they had ideas we did not provide, My main idea was to add a long text field like in sketch V3. As well, since we were moving away from yellow, I made the prompts grey.

While we provided prompts, we wanted to give the user the chance to create their own prompt in case they had ideas we did not provide, My main idea was to add a long text field like in sketch V3. As well, since we were moving away from yellow, I made the prompts grey.

While the look of this was approved, the CEO wanted it to be more obvious that the user could only use one side or the other, not both. I needed visually prove only the prompts or fields could be used at a time.

PART THREE

The QA Testing

QA Testing

Adjusting Mockups

Since this was my first time designing responsive web, I didn't realise until actually testing that some of the spacing I used would not work out as well as on different mobile browser types.

My Approach

I had to readjust some headroom, realign headers, and increase the size of the modals.

AI Limitations:

During testing, the blogs were too long and sterile. The time to correct the tone might have been better used to write original sentences.

My Approach

I requested devs to add extra instructions for input to split text into more paragraphs, and to sound more human.

Cross-Team Communication:

English was not the overseas dev team's first language, and we worked asynchronously. Confusion could result in days to sort out.

My Approach

I used a ticket template and video walkthroughs for more structured and visual feedback to the devs

Scope Creep:

I wanted to develop the design system and work on edge cases like what if users kept pressing next? but we were working on tight deadlines and resources.

My Approach

We decided to disable the "Next" button when the field was empty, encouraging user input. This was to reduce error cases and save time.

PART FOUR

The
Solution

The Solution

Create a blog with prompts!

Users are able to choose either a suggested prompt, or create their own prompt, helping reduce overload when deciding on what to write about.

click to view

click to view

Addressing Brainstorming Pain Point

Brainstorming can be overwhelming, and photographers (unfortunately) don't have photographic memory. Offering three ways to write a blog provides concise guidance.

Addressing Drafting Pain Point

Writing takes a long time due to an almost infinite number of possible drafts. With the custom option, photographers can make a first draft in minutes and can spend time more efficiently tailoring from the generated blog.

Mobile Screens

Since photographers can be on the go, we also made sure to account for mobile-led designs for the blogging process.

PART FIVE

The
Impact

The Impact

Was AI controversial to users?

AI has been seen as a way to cut out the source of creatives but would that include photographers? In the photography communities we brought the idea to, photographers viewed photography and writing as different creative verticals, and we received positive interest on the idea of AI-blog writing.

Our goal became to improve user engagement and lower the barrier to entry for photographers to blog. AI-assisted blogging jumpstarted their blog writing journey, and allowed photographers to onboard themselves more easily in the process.

The key user problem was being unable to make blogs. With the launch of the feature, users were able to make blogs in just a few minutes, which addressed the need for speed as well as adding options for website customization on Focal's website Creator.

After shipping, what happened?

Self serve user activation rate increased from 10% to 70%

The key business goal was to increase more premium subscription plans with access to the admin section. There was an increase in subscriptions at the launch, with more data needed to calculate long-term success, and will require more iterations to polish the product.

The key user problem was being unable to make blogs. With the launch of the feature, users were able to make blogs in just a few minutes, which addressed the need for speed as well as adding options for website customization on Focal's website Creator.

1.6 blogs made for every total user

The key user issue was being unable to blog more efficiently. With the launch of AI-assisted blogging, users were able to make blogs in just a few minutes, which addressed the need for speed as well as adding options for website customization on Focal's website creator.

The key business goal was to attract more users to subscribe to plans with access to the admin section. There was an increase in subscriptions at the launch, with more data needed to calculate long-term success, and will require more iterations to polish the product.

Received Positive Feedback

“Yea baby !!! Whooh. Honestly it’s a massive thank you to you guys. The new Blogging features as well as just the clean easy to use interface has made my life so much better and makes me motivated to make a difference.
Such a massive thank you mate”

— Brenton

Brenton Cox,

Photographer

Learnings

I'm more conscientious of scope creep

I learned to prioritize and not dwell on too many iterations before review, as smaller companies have more leeway to polish after things get pushed.

I see the value of responsive web testing

I learned to integrate device UI for web-responsive mockups, as I only had experience creating theoretical mobile apps.

Reflections for the future

While it was exciting to create an entirely new feature, the reality is that there are limitations to an MVP for the sake of shipping it out and testing it with real users as soon as possible.

Here are some plans for iterations if I had more time:

  1. Include tone input feature somewhere in the flow. This accommodates users who care a lot about the voice of their brand. 

  2. Make the paragraphs individual elements so that users can more easily edit and add headers between the text. Some users in the media brought this up.

  3. Continue to add more prompts, and do design discovery to guide that process.

Thanks for reading!

Thanks for reading!

Want to work together?

Email me at
heidi.a.suen@gmail.com

Say Hello!

Heidi Suen 2024

Richmond, BC

Create a free website with Framer, the website builder loved by startups, designers and agencies.