Already Paying an SEO Agency? One Chart Tells You If It's Working.
Here's something that might surprise you about the calls we get: about half of them aren't from businesses that need SEO. They're from businesses that already have it — an agency, a retainer, a monthly report with a logo on it — and one nagging question they can't shake:
"Is any of this actually working?"
They're not being cynical. They're being reasonable. They've been paying for months or years, the reports arrive on schedule, the numbers in them are always framed as progress — and yet the phone doesn't feel busier, and nobody in the building can independently verify a single claim. The relationship runs entirely on trust, and we've written before about why the agency model tends to end up there.
So this post is the answer we give on those calls, written down. You don't need to be technical. You don't need the agency's permission. You need one chart — monthly organic sessions, thirteen or fourteen months of it — which you can pull yourself in about three minutes, and a guide to the five shapes that chart makes. We'll walk a real example, month by month, including the shape nobody wants to see: the decline that settles into a plateau.
As always, check-it-yourself is the whole point around here — so let's get you the chart first.
Step one: pull the chart yourself (three minutes, no permission needed)
If your business has Google Analytics — and if you have a website, you almost certainly do, even if you've never logged in — the report lives here:
Open Google Analytics → Reports → Acquisition → Traffic acquisition. Filter or find the row for Organic Search — that's visitors arriving from unpaid Google results, which is the thing your SEO retainer exists to grow. Set the date range to the last 13–14 months, and view it by month.
Three settings matter, and each one closes a door that fuzzy reporting walks through. Organic only — total traffic blends in your ads, your social posts, and people typing your name directly, which lets a report claim credit for growth SEO didn't cause. Monthly — weekly views are noise; daily views are chaos; months tell the story. Thirteen-plus months — because you need to see the same month last year, and that comparison, as you're about to see, is the single most honest number in this entire exercise.
(One caveat before you read anything: if the chart shows an abrupt, total collapse — traffic falling to near zero overnight — suspect broken tracking before bad marketing. Analytics tags get knocked off during site updates constantly, which is a plumbing problem, not an SEO problem, and it's checkable in minutes.)
Chart in hand? Good. Here are the shapes.
The five shapes an organic sessions chart makes
Shape 1 — the growth slope. Bars climbing left to right, unevenly but unmistakably, with dips that recover above where they started. This is SEO working. It's what our own published chart from a standing start looks like — jagged, dips included, trending up. If this is your chart, your agency is earning the invoice; tell them we said so.
Shape 2 — the seasonal wave. Big regular swings that repeat at the same time each year — the seasonal-business pattern. Waves are fine. The question is whether each peak beats last year's peak and each trough beats last year's trough. Rising wave: working. Identical wave, year after year: you're paying a retainer to stand still.
Shape 3 — the plateau. Flat bars, month after month, wiggling inside a narrow band. This is the most misread shape, because a plateau at a strong level after years of growth can be legitimate maturity — but a plateau is also exactly what "we stopped doing meaningful work" looks like. The tiebreaker: what's being shipped? A plateau plus a changelog of real pages, content, and fixes is a war being fought; a plateau plus reports full of adjectives is a subscription being collected.
Shape 4 — the cliff. A sudden step down — 30, 40, 50 percent in a month or two — that doesn't bounce back. Cliffs have causes: a Google algorithm update (checkable — Google publishes the dates), a site migration gone wrong, a technical break. A cliff isn't automatically the agency's fault. A cliff nobody told you about is.
Shape 5 — the decline-then-plateau. The one this post is really about: a healthy baseline, then a drop, then months of flat chop at the lower level — the site never climbing back to where it was. This is the shape that quietly costs businesses the most money, because every individual month looks unremarkable ("we're at 1,900, about the same as last month!") while the year-over-year picture is a straight loss. Let's read a real one.
Why smart owners stay stuck on this question for years
Before the worked example, it's worth naming why this question — is it working? — goes unanswered for so long in otherwise well-run businesses, because the mechanics are almost designed for it.
The reporting relationship is one-directional: the agency chooses which numbers you see, over what time window, with what framing — and every choice is individually defensible while the sum of the choices flatters the work. Month-over-month instead of year-over-year. Blended traffic instead of organic. Impressions when clicks are down, clicks when impressions are down. None of it is lying; all of it is curation, and the client has no independent view to curate against.
Meanwhile the switching question feels heavier than it is. Owners tell us some version of: "Maybe it's working and I just can't see it. Maybe leaving resets the clock. Maybe every agency is like this." So the default wins — another quarter, another report, another year — and the retainer becomes the most expensive unread document in the business.
The unlock is realizing the referee already exists. Google Analytics and Search Console are neutral parties: they don't work for the agency, they don't work for us, and they've been recording the whole time. The chart you're about to read isn't the agency's story or a rival's pitch — it's the scoreboard, and you've had a key to the stadium all along.
The worked example: fourteen months, read line by line
Here's an actual monthly organic sessions chart of the kind we see on prospect calls — fourteen months, June 2025 through July 2026. The numbers, so you can follow along:
The baseline (Jun–Oct 2025). Summer 2025 runs a healthy 2,500–3,000 sessions a month. August spikes to 4,664 — probably a specific cause: a piece of content that ran, a local news moment, a seasonal surge. Spikes are nice; baselines are what you live on, and this baseline is solid.
The cliff (Nov 2025). October does 3,031. November does 1,605 — a drop of nearly half, and note that it's far bigger than normal seasonality would explain for most businesses. Something happened here: an algorithm update, a site change, lost rankings on key pages. This is the month that deserved a phone call, a named cause, and a recovery plan. On the calls where we see this chart, we ask what the agency said that month. The most common answer: nothing that anyone remembers.
The plateau (Dec 2025–Jul 2026). And then — this is the part to really see — the site never comes back. Nine straight months chopping between 1,600 and 2,300. Each month, viewed alone, looks stable. January even "pops" to 2,248, which a report can frame as a 29% month-over-month win. But zoom out and the story is unmissable: the business lost roughly a third of its organic baseline in one month and has been living at the reduced level ever since, while the invoices continued at full price.
The year-over-year test — the honest number. Now the comparison the 13-month window exists for. July 2025: 2,853 sessions. July 2026: 2,285. Down about 20%, year over year, same month, same seasonality. June tells the same story: 2,544 then, 1,930 now — down 24%. No month-over-month framing survives contact with this number. Whatever the monthly reports have been saying, the business is paying an SEO retainer and receiving less organic traffic than it had before. That's not "SEO takes time." We're the ones who published the honest timeline, and this site is far past it — the patience window is for climbing out of zero, not for sliding backward from a working baseline.
The three explanations you'll hear — and how to check each one
To be fair — genuinely fair, not rhetorically fair — a chart like that has possible explanations that aren't agency failure. Here's each one, with its verification:
"It's the industry — AI Overviews are eating everyone's clicks." This is a real phenomenon, not an excuse someone invented: Google's AI answers cut clicks roughly in half on queries where they appear, and zero-click searches keep climbing. The check: open Google Search Console and compare impressions against clicks for the same period. If impressions held steady or grew while clicks fell, the headwind story has legs — you're being seen but not clicked, and the right response is a strategy shift toward being the cited answer, which your agency should be proposing unprompted. If impressions fell with clicks, you didn't lose a click-through battle — you lost rankings, and the headwind story doesn't cover it.
"There was an algorithm update." Also frequently true — Google ships several significant updates a year, and publishes the dates on its Search Status dashboard. The check: does an update date actually line up with your cliff? And more importantly: updates punish specific weaknesses — thin content, bad experience, weak authority — which means "an update hit us" is the beginning of an explanation, not the end. The follow-up your agency owes you: hit by what, on which pages, and here's the recovery work. An update named eight months later, with no recovery plan attached, is an alibi, not an analysis.
"Sessions are down but the traffic is higher quality now." Sometimes legitimate — pruning junk traffic is real work. The check: quality claims must cash out in leads: form fills, calls, bookings. If sessions fell 20% and tracked leads rose, that's a defensible trade someone should have told you they were making. If sessions fell 20% and nobody can show you the lead numbers at all, "quality" is doing a lot of unpaid labor in that sentence. (And yes — leads, not sessions, are the number that pays the bills; a sessions chart is the fastest health check, not the final scoreboard.)
The five questions to bring to your next agency call
Print these. Ask them in order. The answers matter less than whether answers exist.
1. "Walk me through our monthly organic sessions for the last 13 months — including November." You've already seen the chart; you're testing whether they have.
2. "What is our year-over-year change, same month, organic only?" One number. If the call reroutes to month-over-month or blended traffic, note the reroute.
3. "What specifically shipped in the last 90 days?" Pages published, content written, fixes deployed — a list, not a philosophy. Retainers buy work; work leaves artifacts.
4. "What caused the drop, and what's the recovery plan?" A named cause with dated evidence and a plan is a partner having a hard quarter. A shrug is a landlord.
5. "Can I see the Search Console data directly?" Not screenshots — access. It's your property and your data. Any hesitation about giving the owner access to their own Search Console is, by itself, most of an answer.
What "working" actually looks like
For contrast, so the target is concrete: working SEO is a chart whose year-over-year months beat their predecessors; a report that shows the real Search Console and Analytics numbers including the bad months — ours are public, dips narrated and all, because that's the standard we think every client of every agency should demand; a monthly call where declines get named the month they happen, not discovered by the client a year later; and a through-line from the work shipped to the leads produced. None of that requires our firm specifically. It requires an agency that treats your numbers like you're allowed to see them — the same code you'd expect from anyone else you pay: pick up the phone, keep your word, shoot straight.
The honest part
Three things, plainly. One chart is a diagnosis-starter, not a verdict — a decline can have causes no agency controls, and the fair move is asking the five questions before making any decision, giving your current agency the honest chance to show you the November email you forgot about. Some plateaus are fine — a mature local site at a strong baseline in a small market can be legitimately near its ceiling, and an honest agency will tell you that and shift the work toward conversion, AI visibility, and new fronts rather than milking a maxed-out channel. And we have an interest here — we're an SEO firm writing about evaluating SEO firms, which is exactly why every check in this post uses your data, pulled by you, verifiable without trusting us or them. The chart doesn't care who's asking.
The bottom line
Four sentences, like always. If you're paying for SEO and can't tell whether it's working, pull one chart — monthly organic sessions, thirteen-plus months — and read the shape: slope, wave, plateau, cliff, or the expensive one, decline-then-plateau. Judge it by the year-over-year number, not the month-over-month framing, and test every explanation against Search Console, Google's own update dates, and your lead counts. An agency with real answers to five plain questions deserves to keep the account; an agency with none has answered a different question. Either way, ten minutes with your own data ends a year of wondering.
Frequently Asked Questions
How do I find my organic sessions in Google Analytics?
In GA4: Reports → Acquisition → Traffic acquisition, then locate the Organic Search channel row and set the date range to the last 13–14 months, viewed monthly. That's unpaid Google traffic — the thing an SEO retainer exists to grow — separated from ads, social, and people typing your name directly. If you can't get in because "the agency has the login," that's finding #1: analytics for your business should live in accounts you own.
How many months of data do I need to judge SEO?
Thirteen at minimum, and here's why that exact number: it puts the same month from last year on the chart, and year-over-year, same month, organic only is the one comparison that seasonality can't distort and month-over-month framing can't dress up. Twelve months shows you a year; thirteen shows you the truth. Three months shows you almost nothing — which cuts both ways, including in a new engagement's favor.
Is a traffic decline always the agency's fault?
No — and the fair checks are in this post. Algorithm updates hit sites for reasons that predate any agency; AI Overviews are suppressing clicks industry-wide even where rankings hold; migrations and technical breaks happen. What is always the agency's responsibility: seeing the decline before you do, naming its cause with evidence, and shipping a recovery plan. The chart forgives bad months; it doesn't forgive silence about them.
What's the difference between sessions, clicks, and impressions?
Impressions (Search Console) = times you appeared in results. Clicks (Search Console) = times someone clicked through from Google. Sessions (Analytics) = visits to your site, from all sources — which is why you filter to organic sessions for this exercise. The three should roughly move together; when they split — impressions up, clicks down — that gap is diagnostic, and in 2026 it usually spells AI Overviews absorbing your visibility, which changes the strategy rather than ending it.
When should we actually switch agencies?
Not on one bad chart — on the response to it. Ask the five questions above; then the decision usually makes itself. Stay if declines were flagged when they happened, causes are named with evidence, work ships visibly, and you have direct access to your own data. Go if the reroute is constant — month-over-month framing, blended traffic, "quality" claims with no lead numbers, or any resistance to you seeing your own Search Console. And whoever you evaluate next, make them show you their own published numbers first; an agency that won't be measured by its own standard is telling you how yours will go.
What if our agency's monthly report says everything is great?
Then the report and the chart disagree, and only one of them is your actual data. Reports can be truthful and still misleading — month-over-month wins inside a year-over-year loss, blended traffic hiding an organic decline, impressions celebrated while clicks fall. The resolution takes ten minutes: pull organic sessions yourself, run the same-month comparison, and bring both documents to the next call. A great agency will welcome the scrutiny — genuinely, that's the tell — because an agency confident in its work wants you looking at the real numbers. It's the other kind that needs the report to stand between you and the chart.
Want a second set of eyes on your chart?
Here's the offer, plainly: pull your organic sessions chart — or don't, we'll pull it with you — and bring it to the free visibility check. We'll read the shape in plain English, run the year-over-year math, check it against Google's update dates and your Search Console, test what ChatGPT, Gemini, and Claude currently say about your business, and tell you straight which explanations hold up — including, when it's true, "your agency is doing fine, stay put." That happens, and we'll say it. Yours whether or not you ever hire us, with a real reply from the founder within one business day. No pressure, no pitch deck.
Get your free visibility check → — mention "second opinion" in the form.
Sources
Google — Managing owners, users, and permissions in Search Console
Search Engine Journal (Pew Research coverage) — AI Overviews cut organic clicks roughly in half where they appear
Search Engine Land (SparkToro study) — Zero-click searches keep climbing