SEO as a product
Or the story of how we grew our organic non-branded traffic by a factor of 10x in less than one year
When I first started working with SEO (Search Engine Optimization) back in 2018 within a dedicated product team I was a bit overwhelmed, to be honest. Almost all readings presented SEO under Marketing, and there I was — having never worked neither with SEO nor Product Management — asking myself 2 challenging questions:
- What the hell is SEO? (my knowledge at the time was quite shallow)
- How to do SEO within a Product Team? (tbh, I didn’t really know how to be a Product Manager at all)
To my pleasant surprise, when looking for answers I found out that SEO had an amazing online community and many niche blogs with quite profound and elaborated content. As such, I was able to answer question number 1. Question number 2 remained a mystery that I would only be able to answer within a few months of experimenting with a multidisciplinary team and fine-tuning based on what worked and what didn’t.
This is the story of how we went from zero focus on SEO to having a dedicated product team aiming at discovery and experimentation and grew our organic traffic considerably.
(I don’t intend to answer the question “What is SEO?” since there are many sources of content that are able to answer this better than I could do in a single text)
A bit of context
I was working at a real estate business heavily based on technology which had been growing considerably over the past years. Due to the complexity and bureaucracy involved in renting or buying real estate in Brazil, we needed to solve this problem at different parts of the users’ journey to many actors involved in the renting/buying process (from the tenant prospect to real estate agents, passing through photographers and even home checkers).
Due to the complexity of our product and rapid growth, we had never before been intentional regarding our SEO. As a matter of fact, when we migrated to PWA back in 2017/18 (I was at a different company still) and aimed at having a native-like experience with our website, SEO was not taken into account and some problems started appearing.
Back when I joined — at the end of 2018 — there were some obvious things that needed to be done. At that time, GoogleBot was not evergreen and so we had some trouble indexing pages heavily coded with JavaScript which made our product look like a blank canvas. We had to tackle both technical (such as internal linking, canonicalization and SSR) and content challenges (such as no-brainer keyword optimizations) so it seemed we had a clear headed roadmap for at least one quarter.
Within 3 months our team was able to address most of these major and crucial issues, and, as these tasks were approaching an end I was once again asking myself: “OK, so how do we do SEO within our Product Team?” since now we would actually start looking much more into optimization rather than fixing obvious problems. Let me walk you through our journey in finding this answer.
The discovery phase
As you might imagine, being a product team means starting with a discovery. Luckily, we already had a comprehensive understanding regarding our targeted audience by analyzing data from a bunch of tools such as Google Search Console and Google Adwords Keyword Planner (paid tools that also gave us great insights were Ahrefs, SEMrush and SimilarWeb) but most quality data from these tools usually reinforced things we were doing — even though they weren’t enough to earn a relevant enough place at the SERP (Search Engine Results Page).
We had a strong hypothesis that we were seen by the eyes of our users and Search Engines as a mere transactional platform, instead of being perceived as a source of reliable content also. This was reflected on the queries and search intent that allowed us to make an impression (usually a “{transactional term} + {some attribute}” such as “rent an apartment in São Paulo”) and by that time we weren’t yet sending strong signals that we were an authority to real estate transactions which meant we weren’t able to compete with bigger and more established players.
What content was possibly missing from our product that might make ourselves turn into a real estate business with authority for transactional and also informational queries? We needed to have a better understanding on what was missing or misplaced which meant that users weren’t able to have the best User Experience within our product (this lack of content also meant that people with similar intents probably were not able to organically land at our product). To be able to know our targeted audience a little better we needed to go to the field and ask the right questions, and so we did.
One result I’d like to share with you: we found out that about 80% of the users usually search for information regarding the region which they intend to move to — the other 20% usually already know the neighborhood from previous experiences — which I know it is a no-brainer, but now backed by data. Out of the users which show interest for more information, more than 60% consult with friends and acquaintances who might have some knowledge on the region. This is an example of a discovery process that lead us to valuable insight: content generated by our own users might be a valuable source of reliable information that is already being searched for in other places.
Of course we had many more insights from different sets of data following similar processes which lead to other product improvements, but the key message here is: even though for SEO the “main user” (emphasis on the quotes) is the Search Engine Bot, all content should be optimized for real people with real necessities. This type of discovery was performed in our team at different times and at different stages of SEO maturity, aiming to answer specific things and gather insights that led us to trace an on-product content strategy.
Even though for SEO the “main user” (emphasis on the quotes) is the Search Engine Bot, all content should be optimized for real people with real necessities.
(Of course we performed other types of discovery aimed at the technical nature of SEO, such as Web Performance, AMP, Internal Linking and site architecture, HTML semantics and other types of optimizations but perhaps these are subjects for another article)
Content is king
I’ve read and heard this phrase many times until I could finally understand the real importance and meaning behind it.
We were at a privileged place as a company where we didn’t have many competitors with the exact same business model. This couldn’t be farther from the truth for SEO. We compete with all types of real estate business for a space at the SERP and this is where the biggest challenges lie. We found out as a company that great content was a way to stand out from the crowd and decided to double down on it at different contexts and different parts of our users’ journey.
As I told you that we were at a privileged place as a company I feel obliged to elaborate. Since we act on the end-to-end rental/buying process and are also heavily based on technology we find ourselves at a very special place where we own an ever-changing-if-needed platform and also have an interface for the entirety of our users journey — from searching to leaving the property, in the case of tenants — which usually takes years. This is actually what differs us from competitors for SEO.
Users proactively provide us with great insights and information when reviewing a property, a neighborhood or a city. Also, the vast amount of data we have allows us to take snapshots of the real estate market. These two sources of data combined with curated and in-house produced content culminate on content that is not searchable anywhere else on the web, and all this informational content strengthens the transactional side of our product since we are perceived as having greater EAT (expertise, authoritativeness and trustworthiness) to talk about real estate in certain regions.
The insights culminated in a bunch of content structures as well as incremental improvements on pre-existent parts of the product. One structure we created that I’d like to mention is dearly and simply called “Living Hub” which consists of an informational structure consolidating data from different sources that illustrates to our users what to expect when moving to any given region.
(I don’t intend to pass this as a straight formula that should be applied to all types of business at all times since I believe every business should have its own strategy and the discovery process might lead to different conclusions for companies at different stages of SEO maturity. What was true for us 6 months ago might not be true 6 months ahead. My intention here is to share the process we took to tackle our SEO within our product team)
How to know if you are heading in the right direction
Measuring things for SEO is quite a challenge. Everyone is doing things slightly differently every day: The users search differently (ever heard of long-tail keywords? and what about voice search?), there are seasonality issues, the competitors are trying different strategies, Google’s constantly experimenting and updating its algorithms, our own product is evolving, etc.
Q: How do you take into account all this variability?
A: You don’t. Just kidding, you actually have to take some premises into account.
A/B testing is not a viable option for SEO, since you might be sending different signals to the Search Engine Bot (and enforcing a specific version may be interpreted as cloaking, as from Google Guidelines, for instance) so you have to be creative. The AirBnb Engineering & Data Science has a pretty comprehensive post on how to perform a type of split testing which allows to validate improvements and isolate for some noise. The Moz blog also elaborates on a similar approach.
Here we adopt this type of approach but with a different engineering architecture to allow for quicker testing more independently from the development team, with remote configurations. We are constantly tracking changes in Clicks, Impressions, CTR and Avg. Position for keyword clusters we identify as being important for our product within internally developed tools as to identify incremental improvements from these experiments. We also monitor individual pages and structures with the same goal.
For example, with this type of approach we were able to validate that transactional pages for regions which had respective informational content performed with up to 37% increment on CTR and with an Avg. Position 21% higher.
(Most of the things we build on the SEO squad are done with this premise in mind: I, as a Product Manager, would like to be able to tweak and adjust these experiments and components quickly, painlessly and have this reflect instantly without the need to deploy. This allows us to upload and expand experiments faster and have a mindset of continuous testing and improvement. Almost all content in structures we created within our scope can be updated seamlessly without the need of further development)
It is much harder to do it alone
In a modern workplace such as this, we are usually incentivized to be pretty autonomous — not only as a team but also as individuals. We need to understand, though, that autonomy doesn’t mean doing things alone and I kind of learned this the hard way.
At the beginning of the discovery process I wasn’t so great at communicating with our main stakeholders and the long-term content and technical improvement strategy was not so well structured which lead to confusion and lack of understanding — this is specially true for technical subjects such as SEO. The lesson was quickly learned and, for a while, monthly reports needed to be sent to our main stakeholders. You need to get buy-in.
I’m happy to say that I have always worked with great and inspiring people and also have had great inspirational leaders that helped guide me through this process by asking me the right questions — which made me rethink more than once the path we were heading as a team and as a company (this is absolutely necessary). All results we’ve had are thanks to the SEO team and people who work alongside us in other teams, such as Content, Performance Marketing, Branding and Public Relations. The more, the merrier.
Conclusion and next challenges
The challenges rely much on the unpredictable nature of SEO: being up-to-date on what’s new in the SEO world, constraints and limitations regarding time to index new pages (and update old content), the time it takes to have some statistically significance for our tests, and others as such; but we also had a great challenge that was tackling our SEO in other parts of our product — and in other products we own as a whole — and improving technical complex aspects such as Performance.
We more than 10X’d the organic non-branded traffic in one year and have been showing steady improvement up until now. We took everything we learned and replicated to other parts of the product with different mindsets adapted to each channel.
SEO might be challenging but with the appropriate methodology it reaps great fruits on the long-term. If there is one thing I’ve learned during this time working with SEO is that “no beats yes”, meaning: don’t screw up, work on incremental improvements, validate, iterate and benefit from it. Don’t be lured by extremely quick wins that may sound a black hat SEO alarm. It pays in the end, even though it might be hard to get the buy-in at the beginning. Start as early as possible and continue as consistently and persistently as possible.
(this content was first published around early 2020 and is now being made public)