Diversified Botanics: Pipedrive rebuilt for B2B and B2C
A new Pipedrive built from scratch, with only validated data, that feeds the Snowflake warehouse
Result
A new Pipedrive in 5 weeks, built from zero
With the ERP customer number as the key and a sync to Snowflake for their own reporting.
- fields per B2B customer from the ERP, with the ERP customer number as the key
- 20
5 weeks after the signature, the team worked in a new Pipedrive built from scratch. The proposal was signed on 24 November 2025, and the team had its logins between 29 December 2025 and 5 January 2026. What you see in no other case: we deliberately did not migrate. The new Pipedrive stands next to the old one and only validated data went in.
Starting point
Where it got stuck
- With 40,000+ unstructured leads and 41,000+ polluted organizations, nobody could trust the data.
- Pipeline stages mixed statuses with activities.
- Snowflake, Shopify, Intercom and the ERP did not flow back into CRM in a usable way.
Approach
What we built
We did not patch the old account. We built a new one with a clean data model and commercial logic for B2B and B2C from day one, then trained the team and supported the go-live. This is the kind of build our CRM implementation covers, with the integrations from custom software and integrations.
- We built a new Pipedrive instance from scratch, without migrating from the old one.
- Pipelines follow the actual deal lifecycle, with three separate logics: B2B enterprise, B2B distribution and B2C.
- Pipedrive data flows to Snowflake, the central reporting warehouse, through the existing sync. Shopify and Intercom are connected too, and the ERP customer number is the key for dedupe and import.
- Calls run through CloudTalk with AI summaries. Next phase: PandaDoc quotes and the ERP link.
See where your growth opportunities are.
The growth scan has three fields. Within 48 hours you know what to fix first.
Outcomes
What it delivered
One view of the pipeline
Management sees pipeline health across B2B and B2C in one place, instead of reports that disagree with each other.
A base that automation can run on
With a clean pipeline structure and fixed data rules, automation runs on something stable instead of patching broken workflows.
Fewer tabs for sales
Reps work from one new system with only validated data in it, and the team learns it from a library of Loom videos and how-to guides about the build.
Scope and next phase
This case is about the rebuild: a new data model, logic for B2B and B2C and a sync to Snowflake. Data flows one way, from Pipedrive to Snowflake. Next phase: PandaDoc quotes and the ERP link.
When do you rebuild Pipedrive instead of cleaning it up?
You rebuild when the data model itself is the problem, not just the data in it. That was the case here. Stages mixed statuses with activities, custom fields were inconsistent, and more than 41,000 organizations carried duplicates. Cleaning would have meant judging every record and still ending up on a model that could not carry automation.
A rebuild is not the cheap option. It costs more than adjusting what is there and it takes longer, because every pipeline, field and permission has to be designed again. We said that before the work started. With a less messy account we would have advised a cleanup and a smaller redesign.
These are the signals we look for before we recommend a rebuild:
- Stages describe what a rep did instead of where the deal stands.
- The same field exists two or three times under different names.
- Reports disagree with each other, so nobody uses them to decide.
- Automation keeps breaking because it has nothing stable to run on.
Can you use Pipedrive for B2C?
Yes, but not by treating every consumer order as a deal. Pipedrive is built around deals that a salesperson moves forward. A webshop order needs no follow-up from a rep, so it does not belong in a pipeline.
At Diversified Botanics we split the two. B2B enterprise and B2B distribution each have their own pipeline logic with commercial stages. B2C started as a pilot with a limited dataset, the last quarter to six months of orders, so webshop orders do not clutter the B2B pipelines. B2C runs through Shopify. Purchase history, product preferences and lifetime value signals land on the contact, so the sales team sees what happens in the webshop without a deal for each order.
That split keeps the pipelines readable. A rep opens Pipedrive and sees conversations to manage, not thousands of closed orders.
How does Pipedrive fit next to Snowflake and the ERP?
Pipedrive sits at the front of the chain: a lead starts in Pipedrive, becomes a customer in the ERP, and the reporting comes together in Snowflake. The existing sync from Pipedrive to Snowflake stayed in place. We only replaced its API token for the new account.
We agreed on purpose not to send data from Snowflake back into Pipedrive for now. So there is no reverse ETL in this build. That keeps the new account small and clean while the team gets used to it.
If you do want warehouse data back in Pipedrive later, decide three things before the first sync:
- Which Pipedrive objects receive data, and which fields are owned by the warehouse so nobody edits them by hand.
- Which key matches a warehouse row to a Pipedrive record. An email address works for people, a fixed external ID works better for organizations.
- How often the sync runs. Scores that change daily do not need a sync every minute.
How do you keep the new account from getting polluted again?
By deciding up front who may change what, and by checking for duplicates on a regular basis. The new data model has fixed rules for contacts, organizations, leads, deals, activities and products.
Duplicates still appear once Shopify and Intercom both write into the same CRM. Surge Dedupe finds them on email, phone and name, and it never merges anything unseen: every merge is reviewed first.
Moving data between systems follows the same logic. Read how we plan a CRM data migration with mapping and a dry run, and when we choose not to migrate at all.
In short
Sales Surge built a new Pipedrive from scratch for Diversified Botanics, which sells botanicals to businesses and to consumers. 5 weeks after the signature, the team worked in it. The new account has separate logic for B2B and B2C and uses the ERP customer number as the key. Pipedrive feeds the Snowflake warehouse, and Shopify, Intercom and CloudTalk are connected. Only validated data went in from the old account, which held 40,000+ unstructured leads.
Questions about this case
What did the rebuild deliver for Diversified Botanics?
A new Pipedrive, built from scratch, that the team worked in 5 weeks after the signature. It has separate logic for B2B and B2C, the ERP customer number as the key and a sync to Snowflake. Next phase: PandaDoc quotes and the ERP link.
Why did Sales Surge not clean up the old Pipedrive account?
Because cleaning would have left the old data model in place. The account held more than 41,000 organizations with duplicates, stages that mixed statuses with tasks, and custom fields that did not line up. Fixing record by record would still end on a weak foundation. Rebuilding costs more and takes longer, and we said so before we started.
Can Pipedrive work for a company that sells both B2B and B2C?
Yes, as long as you do not turn every consumer order into a deal. At Diversified Botanics the B2B side runs on pipelines with commercial stages. The B2C side comes in from Shopify as purchase history and signals on the contact. A deal only exists where a salesperson actually has a conversation to manage.
Does data flow from Snowflake back into Pipedrive?
No. In this build the data flows one way: from Pipedrive to Snowflake, through the sync that was already in place. We only replaced its API token for the new account. We agreed on purpose not to send warehouse data back into Pipedrive for now, so there is no reverse ETL through a tool like Hightouch. That keeps the new account clean while the team gets used to it.
Read on
Read on
- Page: Vita Brevis Fine ArtPortrait studio for family portraits, including with pets (United States)Why read it: 6 → 25+ calls per booker per hour, from 6 to more than 25
- Page: MM MontageRoof and facade cladding, BelgiumWhy read it: 7 users, everyone working in Pipedrive within a month
- Page: Custom software and integrationsSales SurgeWhy read it: Integrations that run on their own, between your CRM and the rest of your stack.
Get started
See where your growth opportunities are.
We set up your CRM and connect it to everything else. What does not exist yet, we build.