The short version
- A middle-funnel conversion rate is a ratio whose denominator is collected upstream. One audit found roughly 92% of raw sessions were not human.
- Traffic up with leads flat is often a click that never happened. On one account organic click-through averaged 0.11% for a month.
- Prove leads still arrive before rewriting a sequence. One lead sync stopped firing when an ad form was rebuilt, with no error anywhere.
- The middle is the real constraint when nothing moves between stages. One CRM audit found roughly 98% of contacts in the stage they started in.
Middle-funnel work acts on a contact who has already arrived: lead scoring, nurture sequences, retargeting, one more email. It gets commissioned when a conversion rate falls, and a conversion rate is a ratio built on counts collected upstream. In one analytics audit we ran, roughly 92% of the sessions feeding that denominator were not human.
What counts as middle-funnel work?
Anything that acts on a contact who already exists rather than on how contacts arrive. It is also the one part of the funnel a marketing team can change without permission.
An upstream failure shows up here first, dressed as an underperforming sequence, because this is where the reporting lives.
Five upstream failures and what each does downstream:
| Upstream failure | What we found | What it does downstream | How it announces itself |
|---|---|---|---|
| Unfiltered bot traffic | ~92% of raw sessions non-human | Inflates every denominator | It does not, until somebody filters |
| Titles losing the click | 0.11% average organic click-through | Traffic looks like a ceiling that is not there | Position reports stay green |
| A sync that stopped firing | An ad form rebuilt with a new identifier | Lead volume falls, no error anywhere | A salesperson asks why it went quiet |
| A daily send cap | A free tier capped at 300 sends a day | Steps begin on different days per person | Only when somebody checks the plan |
| One live lifecycle stage | ~98% of contacts in the stage they started in | Nothing to optimise between stages | Contact counts rise, stage changes do not |
Why does a broken top look like a broken middle?
Because the top supplies the denominator.
We ran an analytics audit on one account expecting a tracking gap. Roughly 92% of raw sessions turned out to be non-human. Filtering had never been switched on, and every ratio quoted from those reports for years sat on it. Bot traffic concentrates on particular paths, so the top-pages report had also been steering what the content team commissioned.
On the same account, positions held steady while organic click-through averaged 0.11% across a month. Nothing downstream can act on somebody who never arrived. More in ranking without clicks.
What breaks a middle-funnel number before anyone touches it?
Usually a feed that stopped. One lead sync ran from the ad platform to a spreadsheet to a notification channel to the email tool. It held until somebody rebuilt the ad form, which issued it a new identifier. The chain was trigger-based, so nothing errored and no failed run appeared in any log. A salesperson found it by asking why things had gone quiet.
The other one is a quota. An email tool’s free tier capped sends at 300 a day, which surfaced mid-campaign rather than during setup. A daily cap turns one send into a queue, so steps timed off the send date begin on different days for different people, and the report compares cohorts that were never comparable.
Rewriting the emails fixes neither.
When is the middle actually the problem?
When nothing moves between stages.
One CRM audit found roughly 98% of contacts sitting in the stage they were created in, with no automation to move anybody. The recommendation was to stop traffic spend until a second stage existed.
Count stage changes rather than contacts. A contact count rises whether or not anything happened to anybody.
A middle-funnel metric is a ratio built on counts collected upstream, and those counts fail quietly.
What order do you check these in?
Upstream first, because each answer changes what the ones below it mean.
Start with whether the traffic is real. Filter known bots and internal addresses out, then rebuild the ratio on engaged sessions. After that, pull click-through alongside average position, since a strong position with almost no clicks is a title problem no redesign touches.
Next, submit a test through every live form and follow it to the destination record. That check would have caught the broken sync, where no log ever did.
Two more before any copy gets touched: the daily cap on the email plan against the list size, and how many contacts changed stage last quarter.
The two audits that belong before any sequence work are how much of that traffic is human and auditing the CRM first.
Frequently asked questions
What is middle of funnel marketing?
The work that acts on contacts who already exist: lead scoring, nurture sequences, retargeting, sales collateral and follow-up. It changes what happens after arrival and cannot repair how they arrive.
Why are my leads not converting in the middle of the funnel?
Check the upstream counts first. The four we find most often are bot traffic inflating the denominator, a click that never happened, a sync that stopped firing, and a daily send cap.
Should you fix the top or the middle of the funnel first?
Verify the top first, because it supplies the numbers the middle is judged on. Where contacts sit in the stage they were created in, the middle is the constraint and more traffic worsens it.
How do you know whether your funnel metrics can be trusted?
Reproduce the counts underneath them: sessions filtered for bots, form submissions matched to destination records, sends checked against the daily cap, and stage changes rather than contacts.
What is a good middle-of-funnel conversion rate?
There is no benchmark worth quoting, because stage definitions and traffic filtering differ at every company. Measure your own movement instead: contacts that changed stage, and from where.
Does lead scoring fix a middle-funnel problem?
Lead scoring re-ranks records you already hold. Where records stopped arriving, because a sync no longer fires, it has nothing new to rank.
Sources
- Chua Network delivery data across 8 client accounts (internal fact bank)
- Chua Network engagement records, anonymized (internal experience bank)
