Soar’s mention rate rose from 22.1% to 56.7% across two tracked agency questions, comparing July 2–31 with August 10–September 8, 2026. Its rank in that sample moved from fourth to first. Parse and Soar share an owner; this is a narrow observational case study, not evidence that Soar is the best agency or that one action caused the change.
| #4 → #1 | 2.57× | 53.3% | +15 points |
|---|---|---|---|
| Our category rank | Growth in our AI mention rate | Answers citing our website, up from 29.4% | Our lead over the next agency |
Soar went from appearing in roughly one in five AI answers to more than one in two. Our mention rate rose from 22.1% to 56.7%, and we held #1 on every recorded day from August 17 through September 8.
We published an agency comparison, expanded our service pages, and answered common hiring questions. In Parse, we could see those pages getting cited, read how AI described us, and find the questions where competitors still led.
We operate both Soar and Parse. This case compares two 30-day windows ending July 31 and September 8, 2026, for two agency questions on ChatGPT Search and Google AI Mode.
The results
More answers mentioned us, and more cited our website:
| Result | July 2–31 | August 10–September 8 | Change |
|---|---|---|---|
| Our category rank | #4 | #1 | Up 3 places |
| Answers mentioning Soar | 22.1% | 56.7% | +34.6 percentage points |
| Answers citing our website | 29.4% | 53.3% | +23.9 percentage points |
| Answers citing our agency comparison | 2.2% | 17.5% | +15.3 percentage points |
| Answers citing our Quora service page | 16.9% | 30.0% | +13.1 percentage points |
Our comparison article went from 3 cited answers to 21, seven times as many appearances in a smaller sample. Our Quora service page grew from 23 to 36.
We tracked mentions to measure how often we appeared, and citations to see which pages the answers used.
Getting on the shortlist
At Soar, we help brands reach relevant conversations on Reddit, Quora, and other community platforms. For our own marketing, we wanted buyers asking AI for an agency to find us.
Those buyers could get a shortlist without visiting our website. Our homepage alone could not explain everything they needed to know.
So we built out the details: how to choose an agency, what working with us involves, and when it makes sense to hire one.
The timeline
Our content history records the April changes. Parse supplies the later discoveries and results.
| When | What we did or observed | Why it mattered |
|---|---|---|
| April 28 | We added our Reddit marketing agency comparison | Gave buyers a shortlist and criteria for choosing a provider |
| April 30 | We expanded our community and Quora service pages | Made our process, buyer questions, and supporting examples easier to find |
| June | Parse surfaced a cited Reddit discussion asking for a Quora agency | Revealed an existing buyer conversation where we were absent |
| July 31 | We ranked #4 with a 22.1% mention rate | Established the earlier measurement window for this case |
| August 17–September 8 | Every recorded daily category rank was #1 | Showed that our lead persisted across 23 daily readings |
| September 7–8 | Stored AI answers cited our pages while recommending us | Let us connect the recommendation to specific source material |
The Reddit discussion was an opportunity Parse identified; we have no verified placement there. The page updates are documented in our content history.
A clearer service page
Our Quora service page answered concrete buyer questions: how to become a recommended answer, reach relevant topics, manage reputation, and connect Quora activity with search visibility.
We also laid out the work: audit the existing footprint, identify questions, publish answers, develop the stronger ones, and measure results. That gave AI answers something specific to say about us.
The page's citation reach grew from 16.9% to 30.0% of sampled answers. On September 8, Google AI Mode described us as:
Best for SEO-driven organic reach.
It then summarized footprint audits, long-form content, and amplification of answers ranking on Google. Beside that explanation, it cited our Quora service page.
Google was using our process to explain why someone should consider us.
If you want to see how AI engines describe your own brand, run a free brand check. It takes a minute.
The agency comparison
Our Reddit marketing agency comparison addressed results, pricing, fit, transparency, AI visibility, and operating risk. We gave buyers criteria for comparing providers and a shortlist to evaluate.
We wrote the comparison and ranked ourselves first. That makes it an owned source, not an independent endorsement.
The article went from appearing in 2.2% of sampled answers to 17.5%. On September 7, ChatGPT Search cited both the comparison and our guide to when a company should hire a community marketing agency. In the same answer, it named us:
Best overall for Reddit/community-driven acquisition.
Both pages appeared as sources beside the recommendation.
Which pages gained ground
The overall rank hid some differences between pages:
Our Quora service page appeared in 23 then 36 answers; our Reddit-agency comparison appeared in 3 then 21.
| Our page | July citation reach | September citation reach | Change |
|---|---|---|---|
| Quora service page | 16.9% | 30.0% | +13.1 points |
| Reddit-agency comparison | 2.2% | 17.5% | +15.3 points |
| Homepage | 2.9% | 8.3% | +5.4 points |
| “Is Quora still worth it?” article | 6.6% | 0.0% | −6.6 points |
| Any page on our website | 29.4% | 53.3% | +23.9 points |
By the later window, more than half of the answers cited our website. Yet the gains were selective. Our broader article about whether Quora was still worth using disappeared from the later sample, while the service page and agency comparison gained reach.
In this sample, the pages about choosing and hiring an agency gained the most. That gave us a useful starting point for deciding what to maintain and expand.
Citation reach means the percentage of answers citing a page at least once. The rows overlap: one answer can cite several of our pages. It is not our share of all citations.
Who else shaped the answers
The source record also showed us that we were competing with publishers as well as agencies.
Parse flagged Growth Marketing Pro's dual role: its website could help shape an answer while its agency competed for the recommendation. We checked the source coverage across our two windows:
| Source website | July citation reach | September citation reach | Change |
|---|---|---|---|
| Our website, soar.sh | 29.4% | 53.3% | +23.9 points |
| growthmarketingpro.com | 27.2% | 36.7% | +9.5 points |
| reddit.com | 47.1% | 31.7% | −15.4 points |
| singlegrain.com | 6.6% | 8.3% | +1.7 points |
Our website became the most frequently cited of these four domains. Growth Marketing Pro also gained reach, so its pages were worth reading alongside our own.
In June, Parse also found a Reddit discussion asking for a good Quora marketing agency. It had been cited ten times during the observation week, but Soar was absent from the page.
That gave us a specific conversation to investigate, with evidence that AI was already using it.
Taking the lead
Here is how the same agencies ranked in July and September:
| Agency | July rank | September rank | July mention rate | September mention rate |
|---|---|---|---|---|
| Soar | #4 | #1 | 22.1% | 56.7% |
| Growth Marketing Pro | #6 | #2 | 19.9% | 41.7% |
| Advant Technology | #3 | #3 | 22.8% | 18.3% |
| Community Boost | #8 | #4 | 16.2% | 18.3% |
| Qadly | #1 | #7 | 39.0% | 15.0% |
| The Community Agency | #2 | #8 | 22.8% | 12.5% |
We finished 15 percentage points ahead of the next agency. We also held #1 on every recorded day from August 17 through September 8. Those daily readings use overlapping windows, so they show a sustained ranking rather than 23 independent experiments.
The rank measures frequency of inclusion across the tracked answers. It does not mean we appeared first in every answer. The dated ChatGPT and Google excerpts show individual recommendations; the table shows our broader progress.
ChatGPT vs. Google
Our strongest improvement came from ChatGPT Search's community-agency answers. We went from appearing in 17.6% to 83.3% of them: a gain of 65.7 percentage points.
| Buyer question | AI model | July mention rate | September mention rate | Gain |
|---|---|---|---|---|
| Best community marketing agency | ChatGPT Search | 17.6% | 83.3% | +65.7 points |
| Best community marketing agency | Google AI Mode | 0.0% | 23.3% | +23.3 points |
| Best Quora marketing agency | ChatGPT Search | 32.4% | 56.7% | +24.3 points |
| Best Quora marketing agency | Google AI Mode | 38.2% | 63.3% | +25.1 points |
We improved in all four combinations. But Google's community-agency answers still omitted us more often than they included us.
Our next step would be to inspect the sources Google uses for the community-agency question. Parse also flagged weaker coverage on comparison and alternative-agency questions in August. Those are specific gaps we can investigate.
What worked for us
The pages that gained reach answered practical buying questions. They explained our work and helped someone choose an agency. We would take that approach again.
Parse let us check the result against actual answers: were we included, which pages were cited, and did the explanation reflect what we offer?
Reaching #1 was a good result. The gap on Google showed us where to keep working.
How we measured it
We compared July 2–31 with August 10–September 8, 2026: 136 sampled answers in the earlier window and 120 in the later one. Both windows cover the same two eligible questions about community and Quora marketing agencies, on ChatGPT Search and Google AI Mode. The category is named Reddit and Community Marketing Services; this sample does not establish leadership across every Reddit-agency question.
We checked our mention totals against the underlying answers and counted each cited page once per answer. We verified our content changes in Soar's repository history and matched the September excerpts to their cited pages. We did not count automated task closures as proof that someone completed the suggested work.
This is our observed progress, not a controlled attribution experiment. We can show what we published, how AI used it, and how our visibility changed. We cannot assign a separate causal lift to each edit or claim revenue from this dataset. The downloadable data contains the calculations and excerpts. Our measurement methodology and guide to mentions versus recommendations explain the metrics.