After Lesson 9, One Step Remains
Lesson 9 was about how to move deals forward after the quote — creating advancing events, using comparable companies from the same industrial cluster as trust anchors, pushing a stalled deal toward the close. When the deal reaches the signed contract, it is a won deal.
But most salespeople stop here: file the contract, add the name to the performance sheet, and open the next item in the follow-up queue.
That step throws away all the information accumulated at a cost of tens of thousands of yuan in sales effort.
A won deal represents more than a single piece of revenue — it is a "buyer profile" that has been validated against reality. Extract that profile and write it back into the filtering criteria from Lesson 1, and next quarter's list will be closer to conversion from the very start.
This action is called the won-deal feedback loop.
I. You Already Have the Best Training Data
Choosing customers in industrial-goods sales is fundamentally a classification problem: this factory will buy, that one will not.
The training data for answering that question is not your intuition — it is the orders you have already closed.
Every won deal validates one thing: this type of factory, under these conditions, when met with this type of sales action, will close.
Every lost deal also validates one thing: a certain type of factory under certain conditions will not close — or the cost is too high to be worth pursuing.
The problem is that this information typically exists in fragmented form inside the salesperson's head, has never been recorded in a structured way, and has never been systematically written back into the judgment of "who to call next quarter."
The result: every quarter's list repeats the filtering logic of the previous quarter — including the judgments that have already been proven ineffective.
We have seen a company selling industrial automation equipment: a sales team of 8, calling approximately 400 factories per quarter on average, closing approximately 12 deals, for a conversion rate of approximately 3%.
The sales director later conducted a review: pulling all 60-plus closed deals from the previous two years and looking at each one — scale, industrial cluster, industry, whether they had export orders, what signal was present at the time of initial contact.
Before the results came in, he expected the won deals to be spread across all types of factories.
After the results came in, he found: more than 70% of won deals were concentrated in metalworking factories in the Yangtze River Delta and Pearl River Delta, with between 100 and 500 employees, and with a record of capacity expansion or new hiring in CNC-related roles at the time of initial contact. Yet this combination accounted for only about 30% of the original list of 400 factories.
Put another way: approximately 280 of the 400 factories they called each quarter had almost no chance of converting from the start — yet salespeople were spending substantial time and calls on those 280 factories.
II. Which Fields to Record for Every Won and Lost Deal
For won-deal information to be reusable, structured fields must be recorded at the time of closing. Not a paragraph of sales impressions — a debrief form with a fixed format.
For every won deal, record these 9 fields:
Basic profile:
- Industry category: What manufacturing sub-sector does the factory's primary product belong to (mechanical processing, electronic components, textiles, chemical raw materials…)?
- Scale range: Annual output value range (below 10 million yuan, 10–50 million yuan, 50–200 million yuan, above 200 million yuan), or employee count range (below 50, 50–200, 200–1,000, above 1,000).
- Industrial cluster / regional affiliation: The industrial cluster or province and city the factory is located in — "Shunde home appliance components cluster" or "Cixi small appliance cluster" is more useful than just "Guangdong Province."
- Export ratio assessment: Has export orders, purely domestic, or uncertain.
Closing milestones:
- Which signal was present at first contact: Capacity expansion record, job posting, customs data growth, winning a bid, new workshop construction — which signal led you to call in the first place.
- Decision-maker level: Owner made the call directly, production/technical manager drove it, purchasing ran the process — who ultimately pushed the deal to closing.
- Decision cycle: How many days from the first call to the signed contract.
Sales actions:
- Most effective advancing event: Was it an on-site demonstration, sharing a peer case, or bringing in third-party validation — which step caused the deal to move forward noticeably?
- Biggest sticking point: Was there a point where the deal almost fell apart? What caused it and how was it resolved?
Lost deals should also be recorded — at minimum, the first 5 fields above plus one additional field: reason for loss (customer delayed, insufficient budget, competing product, needs mismatch, decision chain too long, other).
These 9 fields do not require lengthy descriptions — one sentence per item is enough. Time to complete: approximately 15 minutes.
An 8-person team closing 12 deals per quarter means accumulating 12 forms per quarter — 3 minutes of work per person — but those 12 forms together are the foundation of next quarter's list.
III. Quarterly Review: From Fields to Filtering Criteria
Won-deal forms stored in isolation have no value. Once per quarter, aggregate them — tally the frequency of each field across this quarter's closed deals and find the commonalities.
Steps:
Step one: Tally the industry × scale × industrial cluster distribution of won deals.
Cross-tabulate all won deals across these three dimensions. In most cases, you will find that 60%–80% of closed deals are concentrated in 2–3 cells, not evenly distributed across all dimensions.
Those 60%–80% concentration cells are the sub-segments you should prioritize in the coming quarter.
Step two: Find the shared outreach signals among won deals.
What signals were present when won deals were first contacted? Pull the 5th field from all won deals and aggregate them. If you find that 7 out of 10 won deals had a record of "CNC or digitalization-related hiring" before the first contact, that signal should become a required filter condition for next quarter's list — not an optional reference.
Step three: Compare won deals and lost deals on scale-range distribution.
Won deals concentrated in the 10–50 million yuan range, while lost deals skew toward factories above 200 million yuan — this means large factories have decision chains that are too long and competition that is too intense for your product, and are not worth the effort. Next quarter, move factories above 200 million yuan from the S tier to the C tier and double the resources saved on the 10–50 million yuan range.
Step four: Translate the conclusions into a new filtering sentence.
Write the conclusions from the three steps above as a single sentence (similar to the profile template format from Lesson 1):
Target: metalworking factories in the Yangtze River Delta and Pearl River Delta, 100–500 employees, with recent CNC-related hiring records, purely domestic or with export ratio above 20%. Priority: those with both capacity expansion and hiring signals present simultaneously.
This sentence is the filtering input for next quarter's list.
The full review meeting: 2 participants (sales director + the salesperson responsible for data), 1.5–2 hours, once per quarter.
IV. What Tianxia Gongchang Does in the Feedback Loop
The core value of the feedback loop is: translating the characteristics of won deals into filtering criteria, then running next quarter's list through those criteria again.
If you try to do this step with Tianyancha or 1688, you hit a wall — they classify by registration information, and the commonalities in won deals almost never show up in registration information. The commonalities in won deals are scale range, industrial cluster affiliation, capacity expansion signals, hiring direction — and these must be aggregated from multiple signals.
Tianxia Gongchang covers 4.8 million Chinese physical manufacturing enterprises, each with industrial cluster affiliation, scale tier, export labels, and capacity expansion and hiring signals. After the quarterly review is complete, input the new filtering criteria directly — Tianxia Gongchang outputs a list of qualifying factories, each with signal tags, ready to be scored and ranked as soon as you receive it.
The operating logic of the feedback loop is: the more precise the characteristics of won deals become, the closer the filtering criteria get to the real buyer profile, and the hit rate of the list improves each quarter.
The automation equipment company mentioned earlier, after writing the review conclusions back as new filtering criteria, compressed next quarter's list from 400 factories to approximately 220 — and the number of closed deals rose from 12 to 17. The list shrank by 45%; closed deals actually increased by 40%.
This was not because sales skills improved — it was because every factory called was, from the very beginning, closer to a real buyer.
V. Keeping Lesson 1's Profile Alive — Quarterly Update Rhythm
Lesson 1 produced a one-sentence Ideal Customer Profile — the starting point of the course. That sentence should not be fixed and unchanging; it should be updated once per quarter with won-deal data.
Recommended rhythm:
| Timing | Action |
|---|---|
| Within 24 hours of each signed deal | Fill in the won-deal debrief form (9 fields, 15 minutes) |
| Within 48 hours of each confirmed lost deal | Fill in the lost-deal debrief form (first 5 fields + reason for loss) |
| Last week of the quarter | Aggregate this quarter's won-deal data, run through the four-step review process |
| First week of next quarter | Use the new filtering criteria in Tianxia Gongchang to pull next quarter's list |
| First week of next quarter | Update the one-sentence profile in the sales playbook with the revised version |
Total time investment for this rhythm: 15 minutes per deal, 2-hour review meeting per quarter.
An 8-person team, if it closes 12 deals and loses 20 per quarter, adds up to approximately 8 hours for all debrief forms plus 2 hours for the quarterly meeting — roughly 10 hours total.
10 hours in exchange for a systematic improvement in next quarter's list quality — not relying on any individual salesperson's instinct.
VI. Won-Deal Review Checklist
Won-deal debrief form (fill in after each closed deal)
- Industry category (fill in the specific manufacturing sub-sector)
- Scale range (annual output value or employee count — pick one and keep it consistent)
- Industrial cluster / regional affiliation (to the industrial cluster level, not just the province)
- Export ratio (has export orders / purely domestic / uncertain)
- Signal present at first contact (specific to which roles were being hired for, what was being expanded)
- Decision-maker level (owner made the call directly / technical manager drove it / purchasing ran the process)
- Decision cycle (number of days)
- Most effective advancing event (one sentence)
- Biggest sticking point and how it was resolved (one sentence)
Lost-deal debrief form (fill in after each confirmed lost deal)
- Industry category
- Scale range
- Industrial cluster / regional affiliation
- Export ratio
- Signal present at first contact
- Reason for loss (select one primary reason: customer delayed / insufficient budget / competing product / needs mismatch / decision chain too long / other)
Quarterly review four steps (execute at the end of each quarter)
- Tally won-deal distribution by industry × scale × industrial cluster; find the concentration cells
- Tally won-deal outreach signal frequency; find the high-frequency signals
- Compare won-deal and lost-deal distribution by scale range; adjust tiering weights
- Translate conclusions into new filtering criteria; update the one-sentence profile from Lesson 1
VII. Course Wrap-Up: How the 10 Lessons Fit Together as One Method
This is Lesson 10, and the final lesson in this course. Looking back, the structure of the 10 lessons is a complete feedback loop:
Positioning layer (Lessons 1–2): Before making the first call, first work out "which type of factory to sell to" — use scale, industry, region, and signal conditions to converge the market into a filterable profile, then score and tier factories so that limited sales time is concentrated on S-tier and A-tier accounts.
List-building layer (Lessons 3–4): Translate the profile into combined filtering criteria and build a structured list. At the same time as building the list, solve the "is it a real factory" problem first — fake factories, traders, and shell companies in the list are the largest hidden waste of sales cost.
Outreach layer (Lessons 5–6): Once the list is right, solve the problem of "how to get in." Lesson 5 provides the signal-based cold open script, turning the first sentence from a cold call into an observation with a basis; Lesson 6 maps the factory's decision chain to ensure you are calling the right person.
Timing and qualification layer (Lessons 7–8): Lesson 7 covers timed entry — monitoring 6 demand-window signals and entering within the time window when the factory is most likely to buy; Lesson 8 covers cutting losses early — using the industrial-sales BANT to judge by the second call whether it is worth continued investment.
Conversion layer (Lesson 9): How to move deals forward after the quote. Create advancing events, use comparable companies from the same industrial cluster as trust anchors, and turn the waiting period in a long decision cycle into active advancement.
Review and replication layer (Lesson 10, this lesson): Record every won deal in structured form; extract quarterly review conclusions into new filtering criteria; write them back to the top of the list — so the whole method runs not just once, but self-corrects once per quarter, becoming more accurate with each pass.
The single core assumption underlying all 10 lessons is: finding factory customers is not a matter of luck and individual intuition — it can be broken down into a quantifiable, replicable, and iterable process.
The salesperson's job is to execute this process and, at the same time, turn the results of execution into the input for the next execution. The won-deal feedback loop is the mechanism that makes this happen.
Tianxia Gongchang's position within this feedback loop spans from the profile estimation in Lesson 1, to the list-building in Lesson 3, to the signal monitoring in Lesson 7, to the new filtering criteria coming to life in Lesson 10 — it is the data-layer foundation throughout the method, responsible for turning "the factories you want to find" into "the list you can call today."
4.8 million Chinese physical manufacturing enterprises, each with signal tags, scale tiers, and industrial cluster affiliation — this is the foundation on which the methodology can be put into practice.
The method is not on paper. It is in the filtering criteria that get updated once every quarter.