Organic traffic is up but leads are flat: the five causes, in order

The five causes, ranked by how often they're the real one

Sessions from organic are up 40%, the report is full of green arrows, and sales hasn't noticed anything. Both sides are usually telling the truth, which is why the meeting goes badly.

There are five explanations. They are not equally likely, and — more useful — they are not equally expensive to rule out. Work through them in this order and you'll spend an afternoon instead of a quarter.

Cause one: the traffic is real and the intent is wrong

This is the answer more often than the other four combined. Somebody optimised for volume, the volume arrived, and none of it was ever in the market.

You can settle it in fifteen minutes without a single tool licence.

• Informational traffic is not worthless — it's just slow. It feeds remarketing, brand recall and eventually topical authority. It does not feed this quarter's pipeline, and it should never have been sold as if it would.
• The fix is boring: point the next two months at the middle and bottom of the funnel, and set an explicit split — say 70% commercial, 30% informational — so nobody has this argument again in month nine.
• If your agency can't tell you the commercial share of the clicks they added, they've been optimising a number nobody asked for. That is different from rankings versus traffic, which is the argument one level up.

1. Open the Search Console performance report. Compare the last three months against the previous three, and sort by click difference so the biggest gainers sit at the top.
2. Take the top twenty gaining queries and label each one by hand: buy, compare, learn, or brand. Rough labelling is fine — you are looking for a shape, not a score.
3. If fifteen or more land in *learn*, you have your answer. That traffic was never going to fill in a form this quarter, and no amount of button colour will change it.
4. Now filter the same report to queries containing your commercial modifiers — price, cost, near me, services, company, vs, your product noun. Did that subset grow at all? If it's flat while overall clicks doubled, the campaign grew the wrong half of the site.
5. Finally, check the landing pages behind the gaining queries. Traffic landing on the blog while the service pages sit still is the same finding, viewed from the other end.

Cause two: the leads arrived and the report lost them

Here's the check that separates *no leads* from *leads not recorded*, and it takes half an hour.

Pull every enquiry your business received in the last 90 days from the CRM, the shared inbox and the phone log. Count them. Now pull organic conversions in analytics for the same window. If the business-side number is materially bigger, you do not have an SEO problem. You have an attribution problem, and everything you decide in the next meeting will be wrong.

• A payment gateway or booking widget on another domain. The visitor leaves your host and comes back, the session restarts, and the sale is credited to a referral or to Direct. Google's own documentation on cross-domain measurement is explicit that without it, one person across two root domains is counted as two users and two sessions.
• A thank-you page that isn't tagged, or a form that posts to a subdomain running a different cookie configuration. Self-referrals appear and the original organic source is lost.
• Phone and WhatsApp. A large share of enquiries to Indian service businesses never touch a form at all. If calls aren't tracked, the entire lead channel is invisible to the tool you're using to judge the channel.
• Internal links carrying UTM parameters. Somebody tagged a link in the site's own navigation, and every visitor who touches it gets their session re-attributed to a campaign that doesn't exist.
• Consent banners set to deny by default. Nothing fires until a click that a lot of people never make. Traffic reported through one system and leads through another will disagree permanently.

Cause three: the page answers the question and never asks for the business

The reader arrived, got exactly what they came for, and left satisfied. That's a good page and a bad asset.

This shows up as strong engagement on the article and nothing downstream: time on page fine, scroll depth fine, zero movement toward anything that costs money. It's the natural failure state of blog-led SEO, and it is genuinely the agency's fault when nobody flagged it.

• Every informational page needs one specific next step tied to its own topic, not a site-wide contact link in the footer. A page about GST on agency invoices should offer a GST-ready quote, not "talk to us".
• Name the thing you want them to do and what happens next. "Get a fixed quote in two working days" outperforms "Learn more" because it removes the question the reader was actually stuck on.
• Judge these pages on assisted conversions, not last click. An article that never converts directly but appears in half your winning paths is doing its job — you just have to look for it.
• Check the internal links out of the article. If the only route to a service page is the top navigation, the article is a cul-de-sac wearing a headline.

Cause four: the form works fine on your laptop

Rarer than the first three, and far more damaging when it happens, because it produces a perfect-looking traffic chart and a silence.

Test it yourself, on your own handset, on mobile data rather than office wifi, in the browser your customers actually use. Submit a real enquiry with a real phone number and confirm two things separately: that it reaches the inbox, and that it reaches the CRM. Those are different systems and they fail independently.

• Phone validation that rejects a +91 prefix, or a ten-digit number with a space in it. Silent, ordinary, and it quietly kills a third of submissions.
• A bot-protection widget that fails closed on older Android browsers. The user taps submit, nothing visible happens, and they leave.
• A phone number rendered as text rather than a tel: link, so tapping it does nothing on the one device where tapping it was the whole point.
• A WhatsApp link built for desktop that opens the web client on a phone and asks the visitor to scan a QR code with the phone they are holding.
• The form works and nothing records it. This is the half of cause four that belongs to cause two — you can't tell them apart until call and chat enquiries are being counted.

Cause five: nobody is buying at your price

Sometimes the queries are right, the tracking is right, the pages ask properly, the forms work — and the leads still aren't there. That's the expensive answer, because the fix isn't SEO.

The signature is specific: commercial queries are ranking and getting clicks, visitors reach the pricing page, and they stop. Sales confirms the enquiries that do arrive are real people with real problems and no budget. The channel is working; the offer isn't landing with the segment the channel brings.

Two honest sub-cases. Either the market is smaller than the keyword tool implied — search volume counts curiosity, not wallets — or your price is correct and your page hasn't justified it. The first is a strategy conversation. The second is a copy and proof problem, and it is fixable.

Either way, this is the point where a competent agency should tell you to slow the content spend rather than doubling it. If they never raise cause five, ask why not.

What to ask at the month-six review, and what a good answer sounds like

Four questions. Ask them in this order and the meeting stops being about vibes.

1. "Of the clicks we gained this quarter, what share came from commercial-intent queries?" A good answer is a number and the query list behind it. A bad answer is "traffic is up" — that's the thing you already knew.
2. "How many enquiries did the CRM record from organic, and how many did analytics?" A good answer gives both numbers and explains the gap. Nobody should be surprised by the question at month six.
3. "Which specific pages produced the enquiries we did get?" A good answer names three or four URLs and what they have in common. If the answer is a channel total, no one has looked.
4. "What number are we being measured against, and what was it on day one?" If this can't be answered from a document written before work started, the last six months have no scoreboard and this meeting has no verdict.