Direct answer: Installing GA4 or Baidu Tongji does not mean analytics is accepted. Define the business events and owner first, then handle cookie consent/privacy, then verify data quality. If you only have page_view and no key event, source fields or sales handoff, the report shows traffic—not inquiries.
This report is for owners, marketing/website operations, IT, sales and legal teams at manufacturers in Changzhou and nearby cities. It covers how to accept website analytics, cookie consent, event configuration and data quality. Information is current to 2026-10-30. It is not legal, privacy-compliance or data-audit advice; consent methods, cookie/pixel configuration and retention depend on the business, target market, contracts and counsel.
Acceptance is about whether events answer business questions, not whether the code is installed. A consent banner is not decoration: track after consent. When GA4 and Baidu Tongji disagree, check definitions and filters—do not invent an “accurate number”.
Why can “GA4/Baidu Tongji installed” still fail to explain inquiries?
Because default analytics usually records page views, not form start/submit, phone, WeChat, downloads or sales stage; without events, source fields and an owner, the report cannot answer which channel brings which inquiry.
Many websites only track page_view: they can see visits, sources and pages, but not whether a user started a form, submitted successfully, clicked phone/WeChat or downloaded a brochure—and not whether sales followed up. GA4 key events (formerly conversions) must be marked manually; Baidu Tongji conversions/events also need configuration. Installing code without events leaves you with “how many visits”, not channel/content decisions.
| Layer | What default analytics shows | What is missing | Acceptance question |
|---|---|---|---|
| Traffic | page_view, source/medium, device, region | Business events, source fields, sales stage | Traffic ≠ inquiries |
| Behaviour | Page views, some scroll/click | Form start/submit, phone/WeChat, downloads | Where do users drop off? |
| Conversion | A few automatic events | Key event/conversion goal, success/failure split | Submissions ≠ qualified inquiries |
| Handoff | Email/CRM record | Source field, owner, SLA, stage | Cannot trace back to channel/content |
“What default analytics shows” varies by theme/plugin/configuration; use the actual account and data. Code installation is not acceptance.
Define events first: which actions are worth tracking?
At minimum track product view, form start/success/failure, phone click, WeChat/enterprise-WeChat click, brochure/download and RFQ success; each event needs a name, trigger, parameters, owner and purpose.
More events is not better. Start from the key actions from visit to inquiry, and avoid marking every click as a conversion. Event names should be stable, readable and reusable across pages; parameters should answer source, product, language, page and form type. The table below is a minimum event set for a manufacturer website; adjust to the business.
| Event | Trigger | Key parameters | Use |
|---|---|---|---|
| product_view | Product detail page visible | product_id, category, lang | Product interest/content gaps |
| form_start | First focus on a form field | form_id, page, product | Form drop-off/field friction |
| form_submit_success | Backend confirms success | form_id, page, inquiry_type | Valid submission, not a click |
| form_submit_error | Validation/network/backend failure | error_type, form_id | Technical/UX issues |
| phone_click | tel: link click | page, device, lang | Phone intent |
| wechat_click | WeChat/enterprise-WeChat/QR click | page, channel | Private-domain handoff |
| download | Brochure/whitepaper/drawing download success | file_name, page, lang | Content interest/sales follow-up |
| rfq_success | Quote/sample request success | product, quantity, region | High-intent inquiry |
The event table is an editorial suggestion; GA4 key events must be marked manually, and Baidu Tongji conversion/event capability depends on the account; follow platform documentation and actual tests.
How should GA4 and Baidu Tongji divide the work?
GA4 covers cross-channel/events/key events; Baidu Tongji covers Baidu sources and domestic visits. Both must reconcile with forms/CRM, and one platform cannot replace the other or explain away differences as “error”.
GA4 is suited to event models, key events, channel/page analysis and cross-device leads; Baidu Tongji is more direct for Baidu search, domestic visits and some on-site behaviour. Definitions, sampling, filters, consent state and attribution windows differ, so numbers often disagree; acceptance should record the reason, not force them into one number. Google's key-events documentation defines key events as activities a user marks as important; Baidu Tongji provides visit/source/conversion reports, subject to the account.
| Dimension | GA4 | Baidu Tongji | Acceptance note |
|---|---|---|---|
| Main use | Events/key events, cross-channel, page/product analysis | Baidu sources, domestic visits, on-site behaviour/conversion | Do not use one platform for everything |
| Conversions | Mark key events | Conversion/event by account configuration | Separate success/failure |
| Sources | UTM/referral/organic/ads | Baidu search/ads/external links | Do not guess when UTM is missing |
| Differences | Filters, consent, attribution, sampling | Filters, code placement, definitions | Record the reason |
| Access | Account/property/data stream | Site/account/sub-account | The company must retain ownership |
Platform capabilities/interfaces change; follow the current GA4 key-events documentation and Baidu Tongji account/documentation.
How should cookie consent/privacy be accepted?
Separate necessary, analytics, advertising and external embeds; analytics/ad scripts should load after consent, with refusal/withdrawal and a privacy notice, and consent records; a banner is not decoration.
A common risk is “analytics code runs in the head while the banner is just decoration”. If targeting the EU/UK or processing personal information, the consent order, default state, withdrawal method and privacy policy must reconcile; GA4 consent mode adjusts tag behaviour based on consent state, but it is not legal compliance and does not replace a CMP/privacy policy. China's Personal Information Protection Law requires a legal basis, notice of purpose/method/scope, minimum necessity, and a way to withdraw consent; applicability depends on the business and market.
| Category | Example | Before consent | Evidence |
|---|---|---|---|
| Necessary | Login, form anti-spam, load balancing | May run, with purpose/retention disclosed | Cookie list, privacy policy |
| Analytics | GA4, Baidu Tongji | Load/degrade according to consent state | Banner screenshot, network request, consent record |
| Advertising/retargeting | Google Ads, Baidu ads, Meta pixel | Do not load by default; load after consent | Tag check, consent record, withdrawal test |
| External embeds | Maps, video, live chat | Disclose third party/possible cookies | Embed list, third-party privacy link |
Consent/privacy requirements depend on the target market, business and personal-information type; see GA4 consent mode and China's PIPL; this is not legal advice.
How should data quality be verified?
Check internal traffic, test submissions, duplicates/spam, missing UTM, cross-domain/language, consent refusal and ad blockers; test on real devices/networks, record differences, and never count test data as inquiries.
Analytics acceptance is not about the total number; it is about whether the data can be reproduced. Internal visits, test forms, bot/spam submissions, duplicate submissions, cross-domain loss, overwritten UTM and consent-driven gaps all distort reports. Acceptance should keep test records and filter rules, and mark/exclude test/spam/duplicate leads from sales reports.
| Check | How to verify | Record/handling |
|---|---|---|
| Internal traffic | Company IP/device/account visits | Filter/exclusion rule |
| Test submissions | Submit with a test email/phone | Mark test, exclude from reports |
| Duplicate/spam | Same content/rapid repeats/bot patterns | Dedupe/anti-spam/manual review |
| UTM/source | Ad/email/social links carry parameters | UTM table, landing-page comparison |
| Cross-domain/language | Chinese/English sites, form/payment/download redirects | Cross-domain setup, language parameter |
| Consent refusal | Any requests/data after analytics refusal | Network request, consent record |
| Ad blockers | Browser/plugin/private mode | Record the gap; do not invent data |
| Platform differences | GA4 vs Baidu Tongji vs CRM | Difference note, reconciliation sheet |
The data-quality checklist is an editorial suggestion; filters/consent configuration depend on the platform, market and privacy requirements.
How should access, handover and the 90-day review be accepted?
The company must retain ownership of GA4/Baidu Tongji/tag manager/domain/forms/CRM and at least two admins; handover includes accounts, event table, filters, reports and consent records; the 90-day review is baseline, fix, review.
Analytics accounts sitting with an agency/website vendor is a common handover risk. Acceptance should confirm ownership of the property/data stream/site, tag manager, search platforms, form/CRM and domain; at least two company admins so the team can take over after a departure or vendor change. The 90-day review starts with baseline and technical fixes, then events/sources/forms/sales handoff, and finally separates execution, data gaps and real change.
| Stage | Focus | Deliverable/evidence |
|---|---|---|
| 0–30 days | Account ownership, event table, consent/banner, test submissions | Account list, event table, test record |
| 31–60 days | UTM/sources, form/CRM reconciliation, filters and reports | UTM table, reconciliation sheet, filter record |
| 61–90 days | Baseline comparison, channel/content/product pages, sales handoff | Review report, continue/adjust/stop recommendation |
The 90-day window is an editorial suggestion, not a performance guarantee; without data access or sales feedback, the conclusion must say “not verified”.
Pre-launch acceptance checklist
Before launch, check account ownership, events/key events, consent order, privacy notice, form success/failure, UTM, filters and reports; if any item is missing, do not call analytics “done”.
The checklist can be shared by marketing, IT, sales, legal and the website/agency team. The point is reproducible evidence: who owns the account, where events fire, what happens before/after consent, whether test data is excluded, and whether reports trace back to source and sales stage.
| Check | Pass standard | Evidence |
|---|---|---|
| Account ownership | Company owns GA4/Baidu Tongji/tag/search platforms/forms/CRM | Account list, admins |
| Event table | Product, form, phone, WeChat, download, RFQ defined | Event table, trigger test |
| Key events | Successful submissions/high-intent events marked and verified | GA4/Baidu Tongji screenshot |
| Consent | Analytics/ad scripts load after consent; withdrawal works | Banner, network request, consent record |
| Privacy notice | Purpose, scope, retention, third parties/contact readable | Privacy policy, version |
| Forms | Success/failure/notification/spam/test distinguishable | Test record, backend log |
| Sources | UTM/source fields trace to landing page/channel | UTM table, reconciliation sheet |
| Reports | Traffic, events, forms and sales stage separated | Report sample, difference note |
The checklist is an editorial suggestion, not a substitute for legal/privacy/data-analysis review; configuration depends on the platform, market and contract.
Who should read this, who should not, and what decision does it support?
It fits manufacturers with a website/forms/ads who cannot explain inquiry sources; it does not fit teams without a website, sales handoff, or those wanting analytics to “prove rankings”.
Who should read: owners, marketing/website operations, IT, sales and legal at manufacturers in Changzhou and nearby cities; teams doing a website redesign, promotion, form/CRM or cookie-consent project; leads who must explain “which channel/page brings inquiries” to the owner. Who should not: teams without a website/form/online chat; teams without sales follow-up records who only want visit counts; teams expecting analytics to prove rankings, AI citations or guaranteed inquiries. Decisions supported: GA4 vs Baidu Tongji, which events count as key events, how to order cookie consent, how to attribute forms/phone/WeChat, how to hand over accounts, and what to review in 90 days.
FAQ
Should we install both GA4 and Baidu Tongji?
Depends on the market and team. For Baidu search/domestic campaigns, Baidu Tongji is usually needed; for cross-channel/events/key events/overseas traffic, GA4 is more suitable. If both are installed, accept definition differences and reconcile; do not merge them into one “accurate number”.
Can we launch analytics without a cookie-consent banner?
It depends on the target market, business and personal-information type; there is no universal answer. If consent is required, analytics/ad scripts should load after consent, with refusal/withdrawal and a privacy notice; a banner is only the interaction layer and does not replace the privacy policy, consent records and actual script behaviour. Confirm with legal/platform requirements.
Is a form submission a key event?
“Submit success” can be a key event, but distinguish success/failure, test/real, spam/duplicate; clicking the submit button is not a successful submission. High-intent events (RFQ, sample, quote) can be marked separately, not mixed with ordinary downloads.
Why do GA4 and Baidu Tongji disagree?
Common reasons include code placement/duplication, filters, consent state, attribution window, sampling, cross-domain, ad blockers and bot filtering. Record the reason first, then decide whether to adjust; do not change historical data or invent an average to “align” them.
How should phone/WeChat inquiries be attributed?
Use tel:/WeChat/enterprise-WeChat click events, landing/product page parameters, customer-service channel codes or form source fields; without click events, do not guess from sales memory. The final deal still needs to be traced back in CRM/sales records.
What if the analytics account is held by an agency?
Require transfer of company ownership and at least two admins, including GA4 property/data stream, Baidu Tongji site, tag manager, search platforms, forms/CRM and domain; verify accounts, event table, filters, reports and consent records so the company can take over after a departure/vendor change.
Bottom line
Analytics acceptance is not “is the code installed”; it is whether events answer business questions, consent/privacy reconcile, and data traces back to source and sales. Define key events and account ownership first, then consent and data quality, and review at 90 days—do not treat traffic as inquiries.
Method and limits
This article is public platform documentation, legal text and editorial method, not legal, privacy or data-audit advice for any company, website, account, data or project. It uses no unpublished client accounts, logs or sales data; the event table, consent matrix, data-quality checklist and 90-day plan are editorial, and configuration must be confirmed against the platform, target market, contract and counsel. Platform interfaces/rules change; the latest official documentation governs.
Sources
Information is current to 2026-10-30; each official/primary page governs its own features, terms and version.
- Google Analytics Help, About key events: support.google.com. Definition, marking and use of key events; interface/limits depend on the account.
- Google Analytics Help, About consent mode: support.google.com. Consent mode adjusts tag behaviour based on consent state; not legal compliance.
- Baidu Tongji: tongji.baidu.com. Entry point for visit/source/conversion reports; capabilities depend on the account/current documentation.
- Personal Information Protection Law of the People's Republic of China: cac.gov.cn. Notice, consent, minimum necessity, withdrawal and deletion requirements; applicability depends on the business/market.