Organization & Role
Heap is a digital analytics platform. As their founding documentation strategist, I’m responsible for the user experience, information architecture, content strategy, and all content within the Help Center and developer/API docs.
Project
In May 2020, I launched Heap’s brand-new Help Center. When testing the search experience for Help Center content (via the search bar within the Help Center and popular search engines like Google) I discovered there was a lot of room for improvement, and decided to kick off an SEO improvement project.
Problem
In Google, Help Center content was being outranked by other results when it shouldn’t. At the start of the SEO improvement project, Google Search Console metrics for the site were as follows:
- Average position: 29.3 (results were appearing at the end of the 3rd page 😬)
- Clickthrough rate: 12.3% (only a small handful of users were actually clicking on the results when they found them)
- Total clicks: 3.65k (corresponds to CTR)
- Total impressions: 30.5k (the amount of times a user saw a result)
Additionally, searches within the Help Center itself were not turning up the results that they should. Some notable examples from when I started this project:
- Searches for ‘web install’ did not include the Web Installation guide in the top five results.
- Searches for ‘Salesforce’ did not include the Salesforce integration guide in the top five results.
- Searches for ‘autocaptured events’ did not include the Autocaptured Data doc in the top five results.
I went with the top five results because that’s how many will show in the search drop-down prior to clicking ‘View All’, as seen below. Our data showed that 84% of searchers would click one of the top 5 results, whereas only about 8% would click the ‘View All’ button to see anything beyond that.
Approach
As a relative newbie to all things SEO, I put together a project plan, collected bad search experiences to benchmark success, consulted with experts, completed an SEO certification, and tactically made improvements across the Help Center site, while tracking our success metrics the whole way through.
Gathering Feedback
To gather search results that needed to be optimized, I announced that I was kicking off an SEO improvement project and created a feedback form for team members to file instances of less-than-stellar search experiences that they had encountered or that our customers had reported. The form was structured as follows:
- Search tool (multiple choice: Help Center, dev docs, or Google)
- Exact search terms (text field)
- Expected result (text field)
- Actual result (text field)
- Automatic collection of email addresses as part of completing the form
These form submissions were added into a spreadsheet, which I then updated as I worked my way through the project. Here’s a sample of what that spreadsheet looked like by the end of the project.
Project Planning
As I continued collecting bad search experiences via the form above, I created a standard project plan template based on the one that the Product team uses for their work. This template consists of the following sections:
- Overview: The project owner, stakeholders, key milestones, and project due dates.
- Problem statement: Who is having this problem and why do we care?
- Measuring success: How will we know when we’ve solved this problem? (Includes quantitative and qualitative metrics.)
- Resources: What resources do we need to solve this problem? In this case, I needed access to Google Search Console, budget to book time with a SEO consultant, and time to complete an SEO course.
- Project timeline: Weekly milestones in a table with the status, task, assignee, and due date.
To put together this project plan, I spent time researching the following topics:
- How Google Search Console works, what the different metrics mean, and what metrics to pay the most attention to (mostly reviewing Google’s own content on this subject).
- Best practices for improving SEO for a WordPress site (our Help Center CMS) which lead me to the Yoast SEO for WordPress plug-in and certification.
- How-tos and tips for SEO for Help Center content/documentation hubs, since most SEO guides are geared towards marketing and eCommerce sites.
Beneath the project timeline, I listed out each doc configuration that I planned to update. To check (and share) my understanding of the changes I planned to make, I structured the write-ups with “What is it” and “How should we use it” sections. Here’s a sample:
When finalizing this project plan, I requested feedback from team members who know more about this than me.
Asking the Experts
Our marketing team helped me book some time with an SEO consultant that they worked with. The consultant did a brief audit of the Help Center and sent me some advice for improving SEO, which aligned with the guidance I had learned in my own self-study. This validated that I was on the right track.
I wrote up everything I had learned in the project plan and mapped out a week-by-week strategy. I budgeted two weeks to complete an online SEO for WordPress course part-time, one buffer week for updating the project plan based on what I learned, and 12 weeks to make tactical updates to 25 Help Center articles per week (five articles per day). This would allow me to update our full set of Help Content, around 250 articles, over a 10-week period, with two weeks of buffer time added for if anything went awry.
Throughout this process, I conducted a review of Google Search Console metrics every two weeks to make sure the changes being made were, in fact, improving our SEO efforts. My quantitative goals were to:
- Improve our average position until it was less than 10 (meaning results were displaying on the first page of an average Google search)
- Increase our clickthrough rate, total clicks, and impressions
To better my understanding of SEO for WordPress specifically, and thus finalize the tactical details of this project, I started a free Yoast SEO certification course.
SEO Certification
As part of the Help Center launch, the WordPress agency we worked with to build the site had activated the Yoast SEO plug-in. Though I knew what it was, I had very little understanding of what it was capable of, much less how to leverage it to improve our SEO.
To make the most of this plugin, I spent a couple of hours a day over the course of two weeks completing Yoast’s SEO for WordPress certification. I took notes on what was applicable to the Help Center website, made quick win changes (mostly updating select site-wide settings), and updated the project plan as needed. The course covered the general basics of SEO, SEO specifics for WordPress, and provided a wealth of best practices and how-to information.
Upon completion of the course, I had the revised project plan reviewed by a colleague who owned SEO for our marketing site to validate my understanding and strategy. With his approval, I was off to the races implementing the changes outlined in the project plan.
Tactical Improvements
When planning the project, I made some test updates to estimate how much time it would take me to update all of our docs. I found it took roughly an hour to update five docs, with that estimate varying based on the length and complexity of the doc. Thus, I set about updating five docs per day.
To make it easy to pick back up where I had left off, I worked through the docs changes in chronological order. For quick reference, I created a checklist of things to add/update/check for each doc.
Note: our “In this doc, you’ll learn” and “This doc is for” sections were custom-built by an agency to help readers determine if this was the right doc for them before getting into the details of the doc itself. During my SEO research, I learned these sections have the added bonus of contributing helpful metadata for search engines to decide how to index those articles. Thus, as part of this project, I added and revised these sections for all of our docs.
Here’s an example of what these sections look like in a Heap Help Center doc.
As I worked my way through the project, every two weeks (aka each 20% of the way through) I double-checked the Google Search Console performance metrics referenced in the project plan to make sure they were improving. Exact results are included in the outcome section below.
Outcome
The quantitative results at the end of the project showed a notable improvement in visibility. For comparison, here are the before-and-after metrics (as of the prior 90 days):
| Metric | Start of Project (9/3/2020) | End of Project (11/18/2020) |
| Average position | 29.5 | 13.7 |
| Clickthrough rate | 12.3% | 4.2%* |
| Total clicks | 3.75k | 5.01k |
| Total impressions | 30.5k | 120k |
*A note about clickthrough rate: This value is calculated as the number of clicks on a result divided by the number of impressions. Because the total number of impressions quadrupled, though the total number of clicks only increased by less than half, the clickthrough rate decreased.
Since the primary goal of the project was to increase the appearance of Help Center content in search results, the improvement in average position and total impressions ultimately validated the success of this project, with plans to continually optimize.
Here’s a graph of these metrics, with impressions in purple and clicks in blue.
Qualitatively, I revisited each of the search feedback form submissions filed by the team to re-create the search experience. I found that the majority of these cases were already fixed by the improvements I’d made, with only minor updates needed for more nuanced cases, such as adding a tag for a different way to phrase the same thing (ex. “different count’ or “error in numbers” for “data discrepancies”).
I continue to optimize the search experience via feedback and Google Search Console data to this day.
Reflection
Though improving the average position by 15.8 was very exciting, since the majority of Google searchers rarely look at anything beyond the top three results, there is still work to be done. To this end, I continue to self-educate on SEO and make improvements however I can.
The most common type of feedback received during this project which I was unable to act upon was the desire to be able to enter a search term in the Help Center which would also show search results from the developer/API docs, and vice versa. As of the time of this writing, I am currently in the early stages of an Algolia implementation, which will allow for cross-domain search and a number of other SEO improvements. Stay tuned for an update when this implementation is complete.
