Your Leads Dropped. Here's the Order to Check Things.

A seven-step diagnostic order for a sudden drop in leads, starting with the checks that take ten minutes. Raising budget first is the wrong move.

Work through it in this order. Prove the leads actually stopped arriving rather than stopped being recorded. Split the drop into fewer people looking versus the same people failing to convert. Compare the month to the same month last year before you call it a crisis. Then look for the mechanical failure that produces exactly this shape: a broken form, a page that 404s, a budget that ran out, a suspended profile, a campaign somebody paused. Then competitive and platform change. Then the slow structural causes that only look sudden. The strategy question comes last, after all six.

And do not raise the ad budget first. It is the most common first move, and it is usually the wrong one, for reasons I will come back to at the end.

First, prove the leads stopped arriving and not just stopped being counted

More often than owners expect, the leads did not stop. The record of them did. This check goes first because it is the cheapest one on the list and because every hour spent diagnosing a market that did not change is an hour wasted.

The usual culprits:

• The form. A plugin or theme update, a new required field that fails validation silently on a phone, a spam-protection widget that stopped loading, a mail service that started rejecting the sending domain after a DNS change. The visitor still sees a success message. Nothing arrives.
• The notification. Enquiries landing in a spam folder, filed by a mail rule nobody remembers writing, or sent to the address of somebody who left in March.
• The thank-you page. Change that URL during a redesign and any conversion built on it stops counting. The lead still arrives. The report says it did not.
• The tag. A rebuild that dropped the snippet, a container published without it, a consent banner whose default behaviour changed.
• Call tracking. A number pool that ran out, a lapsed vendor account, a number that stopped forwarding. That last one is the dangerous version, because the calls really did stop.

The test takes ten minutes. Fill in your own form from a phone, on mobile data, with a plausible name. Call your own number from a phone the business does not recognise. Then check every place that lead is supposed to appear: the inbox, the CRM, the analytics property, the ad platform. Wherever the trail stops is your answer.

Then cross-check across systems. Does the drop show up everywhere at once, or in one place? If the phone bill shows normal call volume and the CRM shows half, that is plumbing, not demand. If the ad platform reports the same clicks and fewer conversions, something between the click and the record changed. Our guide to tracking phone leads covers the call side of this in detail, because the phone is where the trail usually goes cold.

Second, separate demand from capture

Two numbers per channel: how many people arrived, and what share of them turned into an enquiry. This one split cuts the search space in half and it takes five minutes.

Do it per channel. An aggregate number hides the case where paid collapsed and organic quietly grew to cover half of it, which is the worst version of this problem because the total looks like a mild dip and the actual failure is total.

Say what you cannot see, too. Some paths have no honest traffic figure. Business Profile views are a platform-reported number you cannot audit, and a direct call from somebody who saw your van has no upstream count at all. A path you cannot measure is a blind spot, not a channel that held steady.

Third, look at the same month last year before treating a trough as a crisis

Month over month is the wrong comparator for most local trades. Compare the same month last year, and the same month the year before if you have it.

Three things distort short comparisons more than people expect. Business days, because a month with more weekend days has fewer working days in it. Holidays, which move around the calendar. And weather, which in a trade is not a distortion at all but real demand. A wet fortnight is a genuine drop in exterior work, and no campaign change will fix rain. In Austin, a run of triple-digit days is real demand for anyone in cooling, and the absence of one is a real absence.

If you do not have a year of history, say so out loud. That is a blind spot, not a clean bill of health. The workable proxies are category search interest over time, what other operators in the trade are seeing, and your own job records if the marketing data does not go back far enough. Pull the year-ago window before the conversation, not during it.

Fourth, the mechanical failures that produce exactly this shape

Read the shape of the line first. Markets slope. Switches drop. If leads went to near zero on a specific day and stayed flat afterwards, you are looking for a switch, and no amount of strategy will explain it.

So pin the date. Get the exact day the line fell, then examine everything that changed within three days either side of it.

• The site. Down, certificate expired, host throttling, or slow enough that mobile visitors leave before it paints. Check the page your enquiries come from, not the homepage.
• A page that 404s. A rebuild, a slug change, a redirect chain, a page deleted by somebody tidying up. If the page your ads point at moved, the ads are still running and still charging you, and sending people to nothing.
• The Business Profile. Suspended, category changed, hours marked closed by a holiday setting nobody removed, or an address edit that triggered re-verification. A suspension produces a cliff, because the profile leaves the map. The rules and the reinstatement process change over time, so check the current guidance rather than what worked the last time this happened. This is one of the reasons Business Profile management is worth someone actually watching rather than setting up once.
• The budget. Exhausted early in the month, lowered by someone being careful, or a shared budget draining into one campaign. Or a card that expired, which stops delivery entirely and sends the warning email to whoever created the account years ago.
• Somebody paused it. A campaign, an ad group, an automated rule that fired on a threshold, a bid strategy change, an ad disapproved. Open the change history first. Every serious ad platform keeps one, and it is the fastest single answer available anywhere on this list.
• The conversion action. If a conversion was deleted, renamed, or demoted to secondary, automated bidding is now optimising toward something else, and delivery changes within days of that.

The change logs worth having open while you do this: the ad platform’s change history, the CMS revision history, the tag manager’s version history, the profile’s edit record, the CRM’s automation history, and the human one, which is asking every person with access what they changed. That last question finds the cause more often than any of the others.

Fifth, competitive and platform change

Now it gets slower and less certain, which is why it is fifth and not first.

A new advertiser bidding on your terms raises the auction price, so the same budget buys fewer clicks. Impression share lost to rank and impression share lost to budget are the two figures that show it. On the organic side, look at rankings by query and by page rather than an average position across everything, because an average hides one page falling off a cliff while forty others hold.

Search results pages change layout, and when more of the screen is taken up by something that answers the query before anyone scrolls, clicks fall while positions do not move at all. That happens often enough that the only reliable read is your own click-through rate at a stable position over time.

Map results vary with where the searcher is standing, so an owner checking from the office is reading one point on a map and not a ranking. We wrote about why a business ranks in one part of town and not another, because this is the check people get wrong most often.

One honest caveat about this whole section: it rarely produces a cliff. Competitive pressure and platform change usually look like erosion. If your line fell off a ledge on a Tuesday, go back to the previous section.

Sixth, the slow causes that only look sudden

Some drops are threshold effects. The cause started months ago and you only crossed the line where it shows up last week.

Review velocity that stalled in the spring. A targeting radius somebody tightened to cut waste. A service page nobody has touched in years while three competitors rewrote theirs. Turnover in whoever answers the phone, which shows up as steady lead volume and falling booked jobs. Or your own capacity: a busy quarter where nobody chased the quotes, arriving now as an empty pipeline.

The test is to plot twelve months instead of two. A gentle slope with a recent step means you have both a structural cause and a mechanical trigger, and fixing only the trigger buys you a few weeks. Rebuilding local visibility once it has eroded is slow work and it is the case local SEO exists for, which is also why it is a poor answer to an emergency.

Seventh, and only seventh, the strategy question

Are we selling the right service, to the right people, in the right places, with an offer the market still accepts? That is a real question and sometimes it is the answer.

It goes last because it is the most expensive thing to act on, the hardest to reverse, and because all six checks above can produce an identical-looking chart for a fraction of the cost. Changing your positioning because a form was broken is a mistake you will be living with next year.

Why raising the ad budget first is usually wrong

The instinct is understandable. Leads are down, ads produce leads, so buy more. Five reasons not to, at this point:

It spends money immediately and diagnoses nothing. If the form is broken, you are buying clicks that cannot convert, and the bill is real either way.

It contaminates the evidence. You have changed a variable in the middle of an investigation, and now you cannot cleanly compare the period to anything.

It resets automated bidding. Platforms describe a learning period after a significant budget or bid change, and the specifics of that change over time, so check the current guidance. Either way, the two weeks after the change are not comparable to the two before.

It treats a capture problem as a demand problem. If capture is broken, more traffic makes the loss bigger, not smaller.

And it is the one action guaranteed to produce a cost while only possibly producing a lead.

The exception, and it is a genuine one: if you already know the budget ran out on the twelfth, refilling it is not strategy, it is repair. That belongs in the fourth section, not this one.

What you should be able to see at any time

This whole sequence takes an hour when the evidence is already there and a week when it is not. The difference is a short list of things worth setting up before you need them:

1. One lead record per enquiry, from every channel, with a source, a date and an outcome. Half of the checks above are impossible without it. The smallest useful version is in our one-page measurement plan.
2. A conversion that fires on one real completion state, tested on a schedule rather than when something looks wrong.
3. An alert on zero. If yesterday produced no leads at all, somebody should hear about it that morning, not at the end of the month.
4. Uptime monitoring on the enquiry page and the form, not only the homepage.
5. Twelve months of history in one place, so seasonality is a lookup and not an argument.
6. Your own accounts in your own name. Ads, analytics, Search Console, the Business Profile, DNS, the CMS, call tracking. If you cannot open the change history yourself, your diagnosis waits on somebody else’s inbox. This is what taking control of your marketing is about, and it matters most on the worst day.
7. A change log a human writes in. One shared place where anyone who touches marketing records the date and what they did. It is the least sophisticated item here and it resolves more of these investigations than any dashboard.

If your leads dropped this week, do the ten-minute version right now, before anything else. Submit your own form from a phone. Call your own number from a number the business does not know. Open the change history on your ad account and read the last thirty days. Most of the time, one of those three ends the investigation.

If you would rather somebody else ran the sequence and told you what they found, that is part of what our marketing services cover, and how it works sets out the order we work in.