My Leads Dropped This Week: Is the Data Just Late?

Your weekly lead count is down, and the daily breakdown shows the damage sitting almost entirely in the last two or three days. If you cut budget over a reporting lag, you break a campaign that was working. If you wait out a real drop, it gets more expensive every day.

Often it is partly late data. Google Ads credits conversions to the day of the click, and they keep arriving for days afterward, so the most recent days are nearly always undercounted. A drop that sits in those days and shrinks when you recheck is lag. A drop spread across the whole week, or one that persists after a few days, isn't lag. It is a real change or a broken pipeline.

Why do the last few days always look low?

Every lead has two dates: when the person clicked the ad and when they filled in the form. Google Ads' default conversion columns file each conversion under the date of the ad interaction that drove it. Someone who clicks on Monday and requests a demo on Thursday becomes a Monday conversion. That conversion did not exist when you checked Monday's numbers on Tuesday.

So recent days are provisional. They fill in backward as late converters arrive. This is what most people mean by the Google Ads conversion delay, and its size depends on how long your buyers take to decide. A pizza order converts in minutes. A B2B demo request might follow a week of comparison shopping. The longer your typical gap between click and form, the emptier the last few days will look on any given morning.

Two things make it worse:

• Recording isn't instant. Even a same-day conversion can take a while to show in the interface. If you check early in the day, yesterday can look thin for reasons that have nothing to do with your buyers.
• Imported conversions arrive when you upload them. If you send qualified leads back from your CRM, they appear only after the upload, and they are credited to the original click date. On a weekly upload cycle, the current week looks empty by design until the file goes in.

Data engineers treat this as normal behavior, not a fault. Data indexed by time arrives out of order. Any count over a recent window is incomplete until the stragglers land. Your lead report is that kind of count.

Is it lag, or did something actually change?

Five checks separate the two, and none takes more than a few minutes.

Look at the shape of the drop. Lag is back-loaded: the early days of the week look normal and the last two or three look thin. A real drop usually starts on a specific day, such as a bid strategy change, a budget cap or a landing page deploy, and stays flat from there. If Monday and Tuesday are as weak as Friday, waiting will not fix it.

Switch the date basis. Google Ads offers conversion columns reported by conversion time rather than by click date. If leads by conversion time look normal, people are still converting. You are seeing click-date attribution fill in, not a collapse.

Note a number and recheck it. Write down last Wednesday's lead count today and look again in two days. If it has climbed, you are watching lag happen. If it hasn't moved and Wednesday is old enough that it should have settled, the leads are not coming.

Count at the source. Pull raw submissions for the same dates from your form backend or CRM. If the CRM has the leads and Google Ads doesn't, the problem is tracking or attribution. If neither has them, they didn't happen.

Check the click side. Clicks and impressions don't wait on a conversion, so they report quickly. If clicks fell alongside leads, the cause is demand or delivery, not timing. If clicks held steady and leads fell on days that should have settled, your conversion rate moved. That is a different diagnosis, covered in whether your cost per click went up or your conversion rate went down.

When is the data lost rather than late?

Waiting fixes lag. It does nothing for a break. For the first few days the two look identical, which is why the source-count check matters.

The usual causes of lost leads:

• Expired connections. CRM and ad platform integrations run on authorization tokens that lapse. When one does, the sync stops outright, often with no error in the dashboard.
• Permission gaps. If the integration user lacks the right to create records or read a field, those records are never written. Nothing is queued to arrive later.
• Syncs that don't backfill. Some integrations only move records forward. Anything missed during an outage stays missed.
• A broken tag. A site release removes or breaks the conversion tag. Leads keep landing in the CRM while Google Ads records nothing.

One cause sits between late and lost. CRM-to-CRM syncs can be throttled by API limits, so records queue and arrive hours later. That is lag, but heavier lag than the ad platform's own.

There is also a third case: leads that exist but are filed in the wrong place. In a data warehouse, a lead loaded before its source or campaign record arrives can land against an "unknown" bucket. The total stays right while the campaign view looks short, until the record is corrected on a later load. CRM routing does something similar. A lead reassigned to a sales stage drops out of the marketing view you happened to check.

The tell is simple. Lag shrinks as you wait. A break stays the same size or grows, and you can usually point to the day it started.

What to do before you touch the budget

Set a settled boundary for reporting. Report performance only on days old enough to have filled in, and label the newest days as provisional. You accept a fixed delay in every report, and in return you get numbers that don't move under you. That is a far better trade than reacting to a figure that will look different on Thursday.

Measure your own lag. There is no universal number. Record day-one lead counts for a few weeks, then compare them with the same days once settled. The gap is your account's real curve, and it tells you exactly how many recent days to distrust.

Reconcile daily, not monthly. Compare CRM lead counts with platform-reported conversions every day. A broken sync then surfaces within a day instead of at month-end, when the missing leads can no longer be tied to anything.

Re-pull a trailing window. If you report from a warehouse or spreadsheet, refresh the last several days on every run rather than appending only the newest day. Otherwise late conversions never make it into the report at all.

Upload offline conversions often. Daily uploads keep recent days from being empty by design. They also give automated bidding its signal while it is still fresh enough to act on.

If the drop survives all of this, it is real. A genuine fall in leads at steady spend is a sudden rise in cost per lead, and the diagnosis from here is broader than timing.

Conclusion

A lead drop that sits in the most recent days is usually the platform filling in, because conversions arrive after the clicks they are credited to. A drop that starts on a specific day, holds after the window should have settled, or shows up in your CRM too is a real problem or a broken pipeline. Look at the shape, count at the source, and wait out only the days your own lag curve says are provisional.

See how Companion answers lead questions from Google Ads and HubSpot data held side by side, with a daily brief on what changed.