Take a SaaS analytics product designed for customer-facing teams and create a space for it in a busy market of Sales/CS tools and traditional analytics tools
Increased average on-site conversion rate from .13% to 1.68% in 7 months
Originally, Sherlock had messaging geared toward Product teams. In Phase 1 of the project, we moved to messaging that focused on actionable insights.
Shortly thereafter, we realized a competitor (who had more features than us) had copied most of our “actionable insights” messaging. Uh oh! We returned focus to our key differentiator — product engagement scoring.
Positioning, Research, Copy, End-to-End Design, Hacky Frontend Coding
As our website reads, Sherlock is the first and only product engagement scoring platform. (For now, all you need to know is that product engagement is a measure of how users and accounts use your product over time, but if you want to dig in you can read Sherlock's Intercom blog.)
Broadly speaking, there are two categories of tools that measure product engagement (or claim to, anyway).
Sherlock scored users and accounts at a deeper level — it's not about the volume of data or masses of tables and charts, it’s about pulling out the most engaged users or the least engaged accounts.
Sherlock sits between those two categories. Its key differentiator is a normalized engagement score that’s digestible.
Why did we believe it matters? Customer stories taught us that SaaS businesses with a try-before-you-buy business model needed a way for the entire organization to be able to access product engagement data at a deeper level. Our spirits were further bolstered by some recent buzz around product-led growth, product qualified leads and a problematic Sales-CS handoff.
First competitor category: Traditional analytics tools like Amplitude and Pendo Full of rich, detailed product data, these tools are too overwhelming to be actionable for “non-data” people.
A few of our current customers said they signed up for Sherlock because their fast-paced Sales and CS teams couldn't get the information they needed from Product. Even worse off were the smaller organizations who didn’t have a data scientist. They really couldn’t get the information they needed.
You should choose Sherlock because: Your Sales and CS teams need targeted insights so they can focus on what matters
Differentiating keywords: actionable, customer-facing teams, integrations, connections
Next up on the competitor list: CRM tools like Vitally and SalesmachineChock full of CRM functionality, they were “actionable” all right. But they used static health scores as a proxy for engagement. Without a deep understanding of how users interact with the product, Sales and CS teams at product-led growth businesses can't find Product Qualified Leads and reduce churn.
You should choose Sherlock because: You need to supplement CRM functionality with data that goes deep enough to give the entire organization the understanding they need
Differentiating keywords: actionable insights, you don’t need another CRM, login/last active is not a proxy for engagement, don’t work in the dark
Hindsight is 20/20
Because Sherlock made product engagement for Sales and CS teams, we decided to position hard against traditional analytics in the first go-around of the redesign. Our biggest differentiator there was our "actionable insights".
However, I didn’t take into account how undifferentiated “actionable insights” was when it came to CRM-y tools. A few months after launching our "actionable" site, we realized Salesmachine had taken our messaging (almost exactly as I had written it!) 😠
The last piece of groundwork: Target personas Previously, Sherlock went after Product people, but the metrics weren't deep enough to make a strong pitch there (although I later explored "be a hero" messaging with that demo).
I teased out a vague idea of a target persona during the research process: Sherlock's ideal customer was someone customer-facing, but that can mean so many things. Sales, CS, founders—all were fair game. The more feedback we got from customers and leads, the more I realized Sherlock's real target demo was a Job to Be Done—the job of taking users from trial to forever.
Developing Brand Voice
Before I jumped into the site redesign, I spent a few minutes (ok, days) deciding how Sherlock would talk to its potential customers.
What type of language would resonate with customer-facing teams? Did we want to be fun? Trustworthy? What balance should we strike between the matter-of-fact language used by traditional analytics and the over-the-top, supercharge-everything tone of CRM software?
I did a few exercises with our team to help us determine where we fell on various scales of personality, voice and tone.
How would we embody our namesake: Sherlock Holmes?
Talking as if we had stepped out of the 1800s would be inauthentic, but I noticed some elements of Sherlockian language that projected the emotions we wanted potential customers to associate with our brand:
I was already tackling a lot, trying to create a new market and all. That’s why I decided to start with an information hierarchy that was fairly standard for a SaaS website. I can change it as the market develops, I thought so I started with this:
Refining the information hierarchySome time after launching, I realized that a revenue lifecycle page wasn't going to cut it and job-title use case pages were still important (even if the focus was on Jobs-to-be-Done). The final iteration of the website had an information hierarchy that looks a little like this:
After several rounds of messaging brainstorms, I landed on a few key themes for the homepage. The next step? Come up with some wireframes and develop the story of the homepage.
We’re competitive people over at Sherlock, so the product designer, the CEO and I decided to have a wireframe competition to see who could come up with the most versions. (Yes, I won! 😀) Here are a few of the contenders:
Homepage 1: "Jobs-to-be-Done"This was a favorite in the beginning.
Earlier competitor analysis had already revealed that using JTBD messaging would help us differentiate from traditional analytics. I created this version to take the focus off the features and put it where product-led growth said it belonged: the benefit to the end user.
What worked: A straightforward, benefit-based approach appealed to fast-paced Sales and CS teams, as they are often looking to just get things done.
Why we chose a different direction: There wasn’t enough information about what Sherlock actually did. Having action-focused Sales/CS teams leave our homepage without a solid understanding of our product was not an option.
Homepage 2: Sherlock-themedI couldn’t resist toying with the brand. Surely there was a way to fit Sherlock’s features and benefits into the Sherlock framework!
I tried positioning the search for engaged users as a Sherlock mystery full of clues. One option expanded on one of our favorite Sherlock quote: “The game’s afoot!” The game of SaaS! Indeed.
What worked: I love the brand. I love Sherlock. It's a relatable story that people already know. By leveraging it, I would be able to connect our brand to character traits like smarts and curiosity. Also, these options evoked the most emotion. It was easy to highlight the anxiety “the other tools” (or lack of tools) gave you when there was a clear story for the reader to follow.
Why we chose a different direction: Form over function. There was no way to use the brand more fully, sure. But we couldn’t make a strong case for the product when everything was forced into a clue or game format. We lost key user benefits like actionability. Not acceptable.
Everyone loves social proof. In fact, most purchasing decisions are made based on reviews and other forms of social proof, especially in SaaS.
Case studies, testimonials, G2 reviews are all ways SaaS businesses use to convince potential buyers their product is best. But what if a homepage did the same thing?
What worked: Our engaged user base of early adopters felt very passionately about our product. They were more than willing to give us testimonials and feedback, which was great.
Why we chose a different direction: We didn’t have a large enough user base or big name logos to make a homepage that was mostly social proof. Someday!
The winning homepage: Putting it all together with Not Superman, SherlockThe wireframe I landed on had elements of all the above options with an added emphasis on our key differentiators.
What we loved about this option:
The pitfalls I was sure to avoid:
The pitfalls I couldn't avoid (and why it seemed worth it):
High on story, low on feature descriptions. This angle had some problems in common with the Sherlock-themed homepage, but that was a risk we were willing to take. Especially with the focus on early adopters and innovators.
The concept of Not Superman, Sherlock was so powerful to us internally that we believed we could rally passionate people around it. They would see the nuts and bolts of the product after signing up for a trial anyway.
Hindsight is 20/20
The more we looked at the page, the more we felt that having specific details about our metrics at the bottom detracted from the overall story. The page was already pretty long as landing pages go and we wanted to lean into the benefit messaging (for focused Sales and CS people!)
We took out the section about the “Sherlock Metrics,” which we felt would appeal more to Product people. Turns out that wasn’t such a good idea.
The design system for Sherlock focused on fun and relatability. While I wanted to incorporate existing brand assets like lineart footprints, magnifying glasses and pipes, I was also painfully aware of the need to remain trustworthy and look like a SaaS analytics product. I decided to pepper traditional-looking product mocks with fun illustrations. Balance!
But that was about as far as I could stretch the illustrations, especially in the section above the fold. It didn’t make sense to put cartoons or stylized illustrations next to messaging that read “We’re not a fantasy. We’re reality.” But this didn’t mean I could get away with stock images of smiling people either — I needed to come up with graphics that conveyed the abstract concepts of analytics.
I cycled through several options with pictures of Sherlock (good for branding, bad for understanding) and a few more with product mocks on a brick background (not eye-catchy enough).
After much Googling, I found an image of a magnifying glass with lines going through it. What if the lines represented certain types of data?
I did a few iterations and landed on this magnifying glass that “found” PQLs, churn threats and top users in random lines of data. That’s exactly what Sherlock does — turns dizzying amounts of data into insights.
Finding the balance between trust and fun
Although we shirked the trends set by the superchargers, we still had to look like we fit in the analytics category and project an air of trustworthiness. Between the bright yellow branding and the downright refusal to use the language used by others in the space, I knew we needed to incorporate design elements that would link us to the SaaS world. You can only be so different, after all! At first, I tried adding some basic floating shapes (like Drift), but it pushed us even farther into the “fun” realm. Adding stylized product mocks on top of the shapes anchored us more in the SaaS world, as the combination of the two is common in SaaS landing page design.
Incorporating social proof Although we didn’t have enough social proof for the social proof homepage, I still needed to incorporate social proof, especially given how new we were. I decided to pair customer testimonials with the value props they talked about. It stretched the social proof farther and linked specific aspects of our product to the people who believed in it.
Sentiment both internally and externally around the new site was generally positive, although some there were still concerns (when is there not room for improvement!)
The initial (round 1) redesign raised the on-site conversion rate from .13% to .93% in 4 months
Concern: Lack of social proofAlthough I incorporated some quotes from our customers, it wasn’t enough to make a strong case that we were trusted by the market as a whole.
To address these concerns, I created a bar with the logos of our customers and a product shot over it.
It was a brilliant way to make it look like we had more customers than we did. (Thank you for the inspiration, Drift!) The truncation of some of the logos by the page boundaries made it seem as if we were overflowing with customers. And no one even noticed we had repeated logos in the bar!
Concern: “Not Superman, Sherlock” was too nuancedOn-page Intercom chats and demo calls indicated that potential customers were confused about what we did. “Not Superman, Sherlock” painted a nice picture but the concept was too nuanced (which was one of our primary concerns in choosing the direction).
The nuanced H1/H2s combined with direct body copy did an A+ job of tying our brand to the thoughts we desired. It did not do such a good job of getting the point across. So I flipped it around — H1s and H2s that focused more on direct value props supported by nuanced body copy.
The switch did increase overall conversion rate, but not enough.
Concern: What about the metrics?As time went on, it became increasingly apparent that people coming to the homepage needed to understand exactly what we did. What made the Sherlock methodology different from other analytics products? I started by adding a reference to our metrics in the overall story, but it wasn't enough.
I had to rework the story to focus on the metrics without losing the benefits customers loved. Customers who “got” us expressed that Sherlock reduced their anxiety — their anxiety about working in the dark, not understanding the data, and having too many tools to go through.
The solution: Data in, differentiated metrics outFeedback from Sales indicated we should continue to highlight our sliders so I made that a focus. From there, I played with a few abstractions of our dashboard and came up with this:
But it took so many scrolls to get through.
“People are going to forget what the story is about by the time they get to the end!”
No problem. I just needed to figure out a way to convey all the same things but in one scroll. Perhaps some more copy throughout the image to explain parts of the picture that needed to be removed.
"Are you sure they took it from us?"
"Yeah, they've been watching us since we started. It's not the first time it's happened."
"Why didn't you tell me Derek 🙄!"
"They keep signing up for our email list too."
To say there was mass hysteria in the office after we realized that one of our biggest competitors had taken our "actionable insights" messaging would be a little overdramatic (although I was definitely a little hysterical).
Without feature parity, we quickly began losing more deals to Salesmachine. And although it was easy to feel upset, the issue clearly wasn’t that someone had taken our messaging, but that our messaging was takeable.
I needed to come up with a unique angle that didn’t overlap with CRM engagement tools but still captured the actionability of our metrics. Luckily, our CEO had already been pondering category creation and how we could attempt to make product engagement scoring its own category. A category we would win by default, being the only ones in it.
According to an article by Craft Ventures, there are a couple things a business needs to do to create a category.
First step: Defining the categoryI did a quick analysis of companies in the space, although this time it was focused on current SaaS superstars (like FullStory and Drift). When FullStory was gaining traction, they used the tagline “Think of it like a DVR for your website.” It was so memorable that, when asked, our entire team still thought of FullStory as a DVR for a website (even though their messaging has long since changed).
It makes sense: anchoring something new in something people already know is a powerful tool. What is the anchor for product engagement scoring? A previous e-book we wrote compared product engagement scoring to credit scoring — relatable enough for me to use it as the H1.
Second step: Convince the world the category is validThe rest of the messaging followed from the H1 and was paramount in convincing the world that what Sherlock had to offer (product engagement scoring) was not only valid, but necessary if you had a SaaS business with a PLG model.
I reframed the value props we already had so they fit a “Like this, only that” structure. In doing so, I was able to relate Sherlock to things people already understood (e.g. analytics) but point out how we were better (e.g. actionable).
Third step: Social proof!The Sherlock site already had social proof, but it needed a visual upgrade. I had to find a way to draw more attention to it!
Our logo bar felt out of place in the new story and there wasn't a clear product shot to overlay on top of it so using that visual paradigm was out. I channeled some of the boldness Drift used on their site and came up with this:
I already suspected the magnifying glass I used in the header section was too much for anyone to parse in the 30 seconds a visitor was likely to spend looking at it. Everyone on the team agreed.
Creating a new above-the-fold layoutSimple product mocks were out — too Salesmachine/Vitally. Abstracted illustrations and symbols were out — too traditional analytics. Conceptual graphics were also out, as I didn’t want to end up back where we were with the magnifying glass.
Back to the big category-creation question: what made product engagement scoring different? Sherlock scored users and accounts at a deeper level — it's not about the volume of data or masses of tables and charts, it’s about pulling out the most engaged users or the least engaged accounts. It’s about transforming the tables of data into (as much as I hate to say it) actionable insights. How could I show that?
Literally. Sherlock pulls the important users and accounts out of data, so creating an image of something being “pulled out” of data could work.
Feedback on the first mock: "There's too much going on"The lineart (which I meant to be a branding element) didn’t work. Is it a table? Is it a ghost? No, it’s confusing.
But the overall concept did well in internal team polls. Everyone liked the idea of “pulling” a key user from a table of data with a de-emphasized computer in the background.
Keeping what worked and discarding what didn't, I came up with this:
When we launch this website we’re going to announce it on our social media saying: We don’t supercharge shit! Even our cell phones aren’t charged here!”
Derek, CEO @Sherlock
seems like you guys've been busy, new marketing site looks great!”
A long-lost customer who returned & even converted a few months later! #theonewhogotaway
Internal feedback was generally more positive than the first iteration, but I was still stuck on which card to show. Which would describe the product best? Which would resonate best with users?
In the end, we decided to turn the concept into a video with rotating cards. Why choose just one!
The round 2 redesign increased the on-site conversion from .93% to 1.68% in 2 months