A solid HubSpot foundation, then an e-commerce migration prepared record by record.
Home > Case study: retail
This case study describes the engagement of a Quebec specialty retailer that sells online: setting up HubSpot Marketing Hub Enterprise on clean data, then cleaning its customer base before migrating its e-commerce platform. More than 97% of customer records were kept, with no duplicate email address.
The retailer wanted a complete, reliable marketing infrastructure built on clean data. Its customers lived in two places: in its management system, and in its online store. As long as the two don't talk to each other, every campaign starts from an incomplete list and every report counts some customers twice. It is a HubSpot engagement of the kind we run as a HubSpot Gold partner agency, of the kind described in our HubSpot implementation method: data first, campaigns next. We do not name the client, and we give no detail that would identify it.
We set up Marketing Hub Enterprise from end to end. Contacts were imported and cleaned from its management system. Two synchronizations keep them up to date: a custom sync with that system, and another with the e-commerce platform. On that foundation, we built the segments, workflows and marketing automations. The order matters: an automation running on dirty data produces the same mistakes, faster.
The retailer changed e-commerce platforms. Before importing anything, we cleaned the customer export against rules written in advance: spam, test accounts and invalid emails excluded, email addresses normalized, each customer's language set, addresses and postal codes corrected, phone number conflicts resolved. Authentication and consent columns were removed from the files: consent does not travel through an export. Every excluded record is listed with its reason, and matching against HubSpot was used to flag customers who were not in it yet.
A week later, the file was ready to load. More than 97% of the customer records received were kept, and more than 95% of addresses were judged valid. No duplicate email or phone number conflict remained, and no authentication or consent column. These are the only figures we publish for this case, because they are the only ones we can document with a dated report. We do not publish marketing performance indicators without their original data.
Share of customer records received that were kept once the cleanup was done
Duplicate email addresses and phone number conflicts left in the import file
Length of the cleanup, from the customer export to the ready-to-load file
The retailer wanted a complete, reliable marketing infrastructure built on clean data. Its contacts came from two systems, its management system and its e-commerce platform, which had to be brought together and kept up to date in HubSpot before launching anything.
We set up HubSpot Marketing Hub Enterprise, imported and cleaned the contacts coming from its management system, built a custom sync with that system and another with the e-commerce platform, then the segments, workflows and automations. We then cleaned the customer base before the online store was migrated.
Once the cleanup was done, more than 97% of the customer records received were kept and more than 95% of addresses judged valid, with no duplicate email and no phone number conflict. The file no longer contained any authentication or consent column.
Because we only publish what we can document with a dated report and its original data. The results of this case that meet that standard concern data quality. An open rate or a revenue figure without its period or starting value does not seem publishable to us.
One week between the customer export and the ready-to-load file. The total length of the HubSpot engagement is not published here: it does not appear in the documents this case is based on.
Sekoia is a lean growth agency that helps small and medium-sized businesses perform better in highly competitive markets. Pioneers in Generative Engine Optimization (GEO) in Quebec.
COPYRIGHT @ SEKOIA – 2026
