Where the reps
started.
Three years of freelance work for startups and small companies — web development, inventory imports, spreadsheet automation, market research. The basics that every later project built on.
00.1Context
The job description was “everything.”
Newly-formed startups and small low-tech companies needed someone who would take any job the founders couldn't or wouldn't. Set up the store. Get the inventory in. Clean the spreadsheet. Find the customer. Write the plan.
For three years, that was the role. Generalist by necessity. Compensation was small; the variety was the point. Each project shipped. Each one taught a primitive that came back later, at scale.
00.2Catalogue · What it actually was
Six categories of work. One ledger.
Loosely sequenced by recurrence. The middle two — inventory ETL and spreadsheet automation — recurred most often, and turned out to matter most.
-
Web development Build · launch · maintain 2016 — '18
WordPress, custom HTML/CSS/JS — landing pages, marketing sites, small business webfronts. Whatever the new venture needed online. Shipped, not designed.
-
Shopify & WooCommerce setups Stand-up · configure · hand off '17 — '18
End-to-end store deployments — themes, payments, shipping, taxes. Configured the platform so a non-technical owner could run it. The first exposure to e-commerce data structures — products, variants, SKUs, attributes.
-
Inventory imports — ERP → web Extract · transform · load · by hand recurring · '16 — '19
The recurring core: spreadsheets exported from an ERP, reshaped into the schema a Shopify or WooCommerce store would accept. Manual transformation, normalization, enrichment, formatting. Done in Excel, repeated for every catalogue refresh. The lesson — that a clean schema is more valuable than the code that imports it — would resurface later, in the same form.
-
Market research Survey · synthesize · brief on demand
Competitor scans, pricing benchmarks, opportunity sizing for whatever segment the founder was about to enter. Outputs: short briefs the founder could read in one sitting and decide from.
-
Data analysis & spreadsheet automation VBA · formula engineering · dashboards '17 — '19
Excel macros and formula chains that turned a weekly two-hour reporting ritual into a one-click refresh. The first time automation paid back its build cost within three months. The reason later pipelines start with a deterministic gate before anything probabilistic.
-
Business plans Frame · model · pitch '17 — '18
Structured business-plan support for founders fundraising or applying to grants. Financial models, market sections, executive summaries. The narrative side of the work — useful later when shipping engineering had to be defended to non-engineers.
00.3The thread
What stuck.
Manually moving inventory between an ERP export and a Shopify import is, structurally, the same work as a five-stage scraping pipeline. The format changes; the operation — extract, normalize, enrich, format, deliver — does not. Three years of doing it by hand built the reflex before there was a pipeline to put it in.
Three years of small jobs also taught the meta-skill that matters more than any one of them: showing up to whatever the work actually was, in whatever shape it arrived, without negotiating its glamour.