First, prove the drop is shared and real
A shared decline has a signature: the channels turn down in the same week, by broadly similar proportions, and the fall appears in the business's own records as well as in the marketing reports. Check all three before diagnosing anything.
Start with timing. Plot each channel's results week by week on one chart and mark the week each began to slip. If the turns line up, something common changed that week; if they are staggered, you are probably looking at separate problems that happen to overlap, and the marketing strategy troubleshooting section or the platform's own hub is the better starting point. Next, compare sizes. A tracking fault usually cuts every source by a similar share, while a demand fall tends to hit the channels that capture existing intent hardest. Google's Explorations let you lay segments side by side, which makes it quick to compare sources across the before and after periods.
Then go to the source of truth. Orders in the shop system, enquiries in the sales inbox, bookings in the calendar. If those held steady while the marketing reports fell, the problem is counting rather than demand, and the first cause below comes first. If both fell, the business has really lost customers and every other cause stays in play.
Diagnosis table
Diagnosis
| Symptom | Check | Likely cause | Fix |
|---|---|---|---|
| Reports fall but orders and enquiries in the back office do not | Tag, banner or consent platform changes in the week of the drop | A tracking or consent change hitting all measurement | Repair the tags or consent configuration; annotate the break |
| Every channel down, worst in search and other intent-led channels | Category search interest and last year's pattern for the same weeks | Seasonal or market-wide fall in demand | Re-phase budget and targets to the demand curve |
| Visits hold up but conversion rate falls on every source | Site releases, outages, error logs and field speed data for the period | A website change, outage or slowdown | Roll back or fix the release; restore speed on key templates |
| Click-through holds, purchases drop at the product or basket step | Price, delivery terms, stock and offer changes against the drop date | A price, stock, shipping or offer change | Restore the offer or state the new value clearly |
| Costs rise in auctions and branded queries show a new name | Auction overlap, rival ad libraries and pricing pages | A competitor entering, cutting price or outspending | Answer on position and proof, not by matching every move |
| A slow slide over months, with branded search flat or falling | When reach and brand budgets were last cut | Brand-building cut earlier, so capture now weakens | Restore protected creation spend and judge it on lagged measures |
Fix, verify and prevent each cause
A tracking or consent change hitting every channel's measurement
Every channel's results pass through the same tags, the same consent banner and the same analytics property. A new consent platform, a tag manager publish, a site template that dropped the container, or a change of banner wording can remove a share of observations from every source at once. Google's consent mode documentation describes how tags change their behaviour according to the consent state they receive, so a configuration change alone can move every report.
Fix: compare the reports with back-office counts, identify the release that coincides with the break, and correct it using the principles in the privacy and consent guide. If the drop began on the day a banner went live, the data drop after a cookie banner page diagnoses the consent states one by one.
Verify: reported conversions and back-office records move back into their usual relationship for a full trading cycle.
Prevent: put tag, banner and consent changes on the same release calendar as site changes, and annotate every one in the analytics property.
A seasonal or market-wide fall in demand
Sometimes nothing is broken: fewer people want the product this month. Weather, school terms, holidays, the economy and news events move whole categories, and the channels that harvest active searches feel it first.
Fix: check category search interest and your own history for the same weeks in earlier years, as the seasonality section of the annual and quarterly planning guide describes, and re-phase spend and targets to the demand curve rather than chasing a flat line.
Verify: your share of the category, measured as your results against category interest, holds steady even while absolute results dip.
Prevent: build the demand curve into the annual plan so that each month's KPIA metric tied directly to an objective. target already reflects the season, and record the assumption the plan makes about demand.
A website change, outage or slowdown that every source lands on
Paid clicks, organic visits and email traffic all arrive on the same templates. A redesign, a checkout plug-in update, a hosting change or an expired certificate can cut the conversion rate (CVRShare of visits or clicks that convert.) of every source at once while visit counts barely move.
Fix: list every release in the window, check uptime and error logs, and read real-user speed by template and device as the landing page speed and mobile guide recommends; roll back or repair whatever coincides with the turn.
Verify: conversion rate by device recovers on the affected templates and the recovery appears across sources, not only in one.
Prevent: run a short purchase or enquiry test on a phone after every release, and keep a dated release log the marketing team can read.
A price, stock, shipping or offer change
Buyers compare the offer, not the ads. A price rise, a new delivery charge, a product out of stock, a lapsed promotion or a changed returns policy reduces purchases from every channel because every channel sends people to the same offer. The tell is that people still click and browse but stop at the product, basket or quote step.
Fix: put the commercial calendar beside the drop date and, where the change was deliberate, make sure the pages explain the value it now buys, following the pricing and packaging guide. Where stock ran out, pause the promotion of unavailable lines.
Verify: add-to-basket or quote-request rates return to their usual relationship with product page views.
Prevent: give marketing advance notice of price, stock and policy changes, and review the offer as a standing item in the monthly report. The guide to planning marketing around capacity and stock sets up that notice as a weekly capacity view and covers pausing and restarting by product rather than across the board.
A competitor entering, cutting price or outspending
A new entrant bidding on your terms, a rival's price cut or a heavy campaign from an established player can take share across search, social and marketplaces at the same time.
Fix: check auction overlap reports, ad libraries, pricing pages and review sites in the way the competitor analysis guide sets out, then respond through positioning and proof rather than reflexively matching price or bids.
Verify: your Share of VoiceYour share of market visibility. and conversion rate stabilise, and prospects stop naming the rival as the reason they chose elsewhere.
Prevent: keep the guide's monthly monitoring routine running between annual passes, so a new competitor shows up as a signal before it shows up as a lost quarter.
Brand-building cut months earlier
This cause is the hardest to see because its effect is delayed. Cutting reach and awareness work rarely hurts the next month; capture channels go on converting the people already primed. Over following quarters fewer new people enter the market already knowing the name, branded search flattens, costs per result rise, and every capture channel weakens together.
Fix: check when creation spend last fell and whether branded search and new-customer share have drifted since, then restore a protected share as the brand versus performance guide describes.
Verify: branded search and new-customer share stop falling over the following quarters, and capture costs level off.
Prevent: write the creation share into the plan with its own lagged measures, so a short-term saving cannot quietly empty the pipeline.
A prevention routine for shared declines
The causes on this page share one feature: each is visible somewhere outside the channel dashboards before it shows up as a decline. A routine that watches those places turns a mystery into a dated, explainable event.
Keep a single change log covering site releases, tag and consent changes, price and stock decisions, promotions and budget moves, and require an entry before anything ships. When results turn, the log answers the first question in minutes. Weekly, reconcile the marketing report's conversion count with the back-office count, so a measurement break is caught within days rather than discovered at month end. Monthly, read results against the seasonal curve from the annual plan instead of against last month, and read competitor signals alongside your own figures. Quarterly, look at the slow indicators that only move over time: branded search, the share of customers who are new, and capture costs. A gradual drift in those is the earliest warning that creation work has been starved. Finally, agree in advance who calls a shared-decline review and who attends: marketing, whoever owns the site, finance and sales each hold part of the evidence, and the diagnosis is faster with all of them in the room than with each channel owner defending their own numbers.
A homeware retailer's bad month
An online homeware retailer saw paid search revenue fall 28%, paid social 31%, email 25% and organic 30% in the same month. Because the falls were similar in size and began in the same week, the team stopped adjusting bids and looked for a shared cause. Back-office orders were down only 9%, so most of the fall was counting: a new consent banner, released on the 4th, had shipped with analytics consent never updated after a visitor accepted. Fixing the update closed most of the gap within a fortnight. The remaining 9% matched a delivery charge introduced the same week; basket-to-checkout rate had dropped from 52% to 44%. Reinstating free delivery above a basket threshold of £60 restored it within a month, and both changes went into a shared change log.
Common mistakes
Optimising each channel separately when they all fell together. Treating a reporting fall as a sales fall without checking the back office. Comparing with last month instead of the same weeks last year. Matching a competitor's price before checking whether buyers were leaving over price. Reading a slide that began after a brand cut as bad luck.
