Problem → Process → Measured Outcome
Each case study is structured around the client challenge, SDES production workflow and the business result achieved.
How Real Businesses Solved Data Problems by Outsourcing to SDES India
Real client outcomes are the most honest evidence of what a data entry outsourcing partner actually delivers. Every case study on this page represents a genuine client engagement — the challenge, the solution, the process and the measured result. Our clients include eCommerce businesses, recruitment agencies, healthcare providers, law firms, real estate companies and many other business types who came to us with specific, practical outsourcing challenges and left with reliable, scalable solutions.
These case studies cover the most common outsourcing scenarios our clients face: building a product catalog at launch scale, processing a large medical record backlog, standardising a candidate CV library for ATS upload, maintaining a CRM database for an active sales team, and dozens of others. If your situation is similar to one described here, contact us with your specific requirements — every engagement begins with a free pilot so you can verify quality from your own data before any production commitment is made.
Client names are kept confidential. All other details — volumes, timelines, outcomes — reflect actual project parameters.
Jump to an industry
Amazon Catalog Built from Supplier Files — 11,400 SKUs Across 9 Categories
Client: Multi-Brand Consumer Goods Seller, USA
A US-based seller managing nine product categories had accumulated product data across 34 supplier Excel files, 12 PDF spec sheets and three legacy platform exports — all in different column structures, with different naming conventions and inconsistent attribute vocabulary. Internal staff had spent two months attempting the catalog build without completing it. Upload error rates on the partial catalog submitted to Seller Central were running at 22%.
We began with a source data audit mapping each supplier file to the required Amazon category template, identifying which fields were consistently available and which would require research or exception flagging. Category-specific required attributes for all nine Amazon categories were confirmed before any entry began. A pilot batch of 200 products across three categories was processed and reviewed with the client before full production. The full catalog was processed in batches of 1,000, with flat file validation after each batch. Parent-child variation families were built for 680 configurable products. Image URL structures were prepared from the client's S3 bucket file naming convention.
The completed catalog of 11,400 SKUs was delivered across 14 business days. Upload error rate on the final flat file submission was 0.4% — confined to three categories where Amazon had updated category requirements during our production period. All 22% error items from the client's prior attempt were corrected. The client's first catalogue review confirmed every variation family was correctly structured.
" The difference between what we had tried to do internally and what came back from SDES was significant. We had the full catalog live on Amazon in three weeks from when we sent the files. — eCommerce Operations Manager
Shopify DTC Store Launch — 2,400 Apparel Products with Full Variant Setup
Client: Fashion Brand, United Kingdom
A UK fashion brand launching their direct-to-consumer Shopify store had product data across design team specification sheets, a wholesale CSV file and a legacy Magento export from an earlier website. The data was not formatted for Shopify and contained no variant structure — all sizes and colours were listed as separate flat products. Images were named by internal style codes that did not match the wholesale CSV SKUs. The brand had a hard launch deadline for a PR campaign.
We mapped the three source formats to a unified product master, resolved the SKU-to-style-code image naming discrepancy, built the Shopify variant structure with size and colour options from flat products, assigned collection memberships, populated metafields for material composition and care instructions, completed SEO title and meta description fields for every product and sequenced gallery images in the approved order. Products were delivered in batches of 300 with client review built into the schedule. Final delivery was three days ahead of the launch date.
All 2,400 products were live on Shopify by launch day. The brand's development agency confirmed the import was the cleanest catalog file they had received from a client in two years. Zero re-imports were required. The launch PR campaign ran as planned.
" We had already moved the launch date once because of the catalog. Having SDES take it off our plate and deliver ahead of schedule meant we could focus on the campaign itself. — Head of Digital, Fashion Brand
Magento Catalog Audit and Cleanup — 8,300 Products, 5 Years of Accumulation
Client: Industrial Supply Distributor, USA
An industrial supply distributor had grown their Magento catalog to 8,300 products over five years across 14 product managers and three separate supplier onboarding projects. Attribute vocabulary had drifted — the same material type appeared as "stainless steel", "SS", "304 SS", "stainless" and "SS304" across different product records. Category assignments were incorrect for 19% of the catalog. 620 duplicate SKUs had accumulated from repeated supplier imports. The search and layered navigation on the store was producing incorrect results.
We ran a full catalog audit, categorising quality issues by type and frequency before any correction work began. A standardisation vocabulary of 680 approved attribute option values was confirmed with the client's product team. Attributes were standardised across all 8,300 records, duplicate SKUs identified and merged according to agreed merge rules, category assignments corrected using the current Magento category tree and a corrected catalog import file prepared. The change log documented every modification made at the record and field level.
Post-import, layered navigation filter results were correct for all 14 attribute filter types. Category page product counts were accurate. Search results for material and specification searches returned the correct products. The product manager team reported that supplier file imports now produced far fewer quality issues because the vocabulary standard was documented.
" The navigation on our store had been broken for so long that our sales team had stopped using it. Now it works correctly and customers can actually find products. — Product Manager, Industrial Supply
EHR Migration Data Entry — 4,200 Active Patient Records into athenaHealth
Client: Multi-Location Medical Practice, USA
A US medical practice with four locations was migrating from paper charts and a legacy practice management system to athenaHealth. The legacy system could export a partial data set, but patient health histories, allergy records, current medication lists and secondary insurance details were in paper charts that had not been digitised. The go-live date was fixed by the EHR vendor's implementation schedule.
We established a HIPAA-conscious production workflow with NDA executed before any patient files were shared, access restricted to three assigned operators and PHI transferred via encrypted file share. Patient records were processed in priority order — active patients with upcoming appointments first, then other active patients. Each record was entered into athenaHealth's patient module with demographic fields, insurance details (primary and secondary), health history, current medications, allergy records and any documented chronic conditions. Incomplete or conflicting information was flagged in a review list for clinical staff rather than entered with assumed values.
All 4,200 active patient records were entered and validated before the go-live date. The practice's health information manager reviewed 120 flagged records, all of which had legitimate gaps in the paper source that needed clinical confirmation. Post-migration, the practice reported a 94% reduction in chart pull requests during the first month — records that previously required staff to locate and retrieve paper charts were now accessible in the system.
" We were worried the migration would take longer than the vendor's implementation window. SDES completed the patient records ahead of schedule and the exception list was specific enough that our clinicians could clear it quickly. — Practice Administrator
Patient Survey Data Entry — 3,800 Clinical Questionnaires for Statistical Analysis
Client: Healthcare Research Organisation, Australia
An Australian healthcare research organisation had collected 3,800 patient-completed paper clinical questionnaires across two research sites over 18 months. Questionnaires covered 42 fields including validated scale instruments, multi-select symptom checklists, demographic items and free-text comment sections. The research timeline required the complete dataset for statistical analysis within four weeks.
We confirmed the coding framework for all scale instruments and multi-select fields with the research team before production. Each questionnaire was transcribed field-by-field with Likert scale directional coding applied consistently, checkbox responses coded to the agreed numeric scheme and free-text responses captured verbatim. Partially completed questionnaires were transcribed for all present fields with missing items coded according to the research protocol rather than left blank. 114 questionnaires with ambiguous markings were separated into a review list with the specific ambiguity noted for researcher review.
The complete dataset of 3,800 records was delivered within 12 business days — ahead of the four-week deadline. The research team's data cleaning check found an error rate of 0.3% — within the threshold the study protocol had defined as acceptable for manual entry. The free-text comment fields were delivered as a separate coded file structured for qualitative analysis.
" The exception list was exactly what we needed — specific enough to resolve quickly with the research team. The turnaround was much faster than we expected for this volume. — Research Data Manager
Grantor-Grantee Index Build — 14,000 Deed Records Across Three California Counties
Client: Title Company, California USA
A California title company was expanding their title plant coverage to include three additional counties and needed a grantor-grantee index built from county recorder document image archives covering 2005 to 2023. County recorder portal access was available but required manual field-by-field capture — no bulk export was available. The title plant software required a specific import format.
We established county-specific field capture templates for each of the three counties — each with different recording formats, different instrument type classifications and different handling of trust and corporate entity ownership entries. Production was structured in date-period batches with coverage tracking confirming completeness for each quarter before proceeding. Grantor and grantee names were entered in the format the title software's search function required. Legal descriptions were captured verbatim. All instrument type classifications followed the title company's defined vocabulary.
14,000 deed records were indexed across the three counties and delivered in the title plant import format over six weeks. Coverage was confirmed as complete for every quarter in the date range. The title company's production manager reviewed a sample of 200 records from each county and confirmed naming format and instrument type classification were consistently applied.
" We had been quoting turnaround times to clients that we could not always meet because our plant coverage was incomplete. This project fixed that problem for three counties. — Title Plant Manager
Accounts Payable Invoice Processing — 400 Invoices per Week, Overnight Turnaround
Client: Logistics Company, USA
A US logistics company was processing 350–450 supplier invoices per week manually — the AP team of two was spending 60% of their time on data entry rather than on exception management, payment approvals and supplier reconciliation. Invoice formats varied across 180 suppliers. The company used NetSuite with a customised chart of accounts. Turnaround requirements were tight — invoices received by 5pm EST needed to be in NetSuite for morning review.
We built a GL account coding map from the client's chart of accounts, vendor master and invoice category history before production began. Production runs processed batches received each day and were returned by 9am the following morning. Header fields (vendor, invoice date, invoice number, due date, currency) and all line items (description, amount, quantity, unit, GL code, cost centre) were entered. PO matching was flagged where available. Duplicate invoice numbers against the same vendor were caught and separated before entry. Invoices with missing information or mismatched amounts were flagged with specific notes in a daily exception log.
The AP team's data entry time was reduced from 60% to under 10% of their working hours. Invoice processing turnaround consistently met the overnight SLA. The daily exception log averaged 12 flagged items per week — all with specific, actionable notes — giving the AP team a focused review list rather than a full audit requirement. After 90 days the client expanded the arrangement to include a second entity.
" Our AP team was drowning in typing. Now they actually do AP work — chasing approvals, managing payment terms and dealing with suppliers. That's the job they were hired for. — VP Finance, Logistics Company
Closed Matter Archive Indexing — 26,000 Pages Made Searchable for Due Diligence
Client: Litigation Law Firm, USA
A US litigation firm was undergoing due diligence for a merger. The acquiring firm required searchable access to a decade of closed matter files — 26,000 scanned pages of correspondence, pleadings, orders and exhibits stored as sequentially numbered TIFF files with no document-level metadata. Locating any specific document required opening individual files. The due diligence deadline was eight weeks away.
We established a document type taxonomy with the firm's litigation support team covering 24 document types used in the file types present. Each page image was reviewed, classified by document type, assigned a matter number from the visible case reference, dated from the document date field, named according to the agreed convention and indexed in the document management system import format. For multi-page documents, document groups were identified and the page range captured as a single index entry. Pages where legibility prevented reliable classification were flagged for attorney review.
The complete index was delivered in five weeks — three weeks ahead of the due diligence deadline. The acquiring firm's legal team was able to search the archive by matter, document type, date range and keyword within the text layer created during processing. The firm's managing partner noted that the indexed archive had value beyond the merger — it had made their own historical research significantly faster.
" We had twenty years of files that we'd never been able to search. The due diligence requirement finally forced us to address it and the result was better than we expected. — Managing Partner, Litigation Firm
Shipment Record Entry and BL Processing — Daily 200+ Consignments, Singapore Operations
Client: Freight Forwarder, Singapore
A Singapore freight forwarder processing 200–280 daily shipments was using a TMS that required manual shipment record creation — the system had no EDI integration with their carrier partners. Operators were creating records for each consignment from bill of lading PDFs and email confirmations. Data entry was consuming three full-time staff positions and creating a bottleneck that delayed departure confirmations to importers.
We established a daily production workflow with two dedicated operators trained on the client's TMS through a two-week supervised pilot before live production. Each daily batch of BL PDFs and email confirmations was processed overnight (leveraging the India–Singapore time zone) and returned before the Singapore business day began. Shipment records covered all carrier, routing, commodity, container, weight, dimensional and party fields. Daily exception logs flagged incomplete BL documentation and booking reference discrepancies for the Singapore operations team to resolve before vessel cut-off.
The Singapore operations team's daily shipment record backlog was eliminated within two weeks. Three staff members were redeployed to customer service and carrier management. Importer departure confirmations were dispatched by 8am Singapore time on each sailing day — previously this had often not happened until mid-afternoon. The client extended the arrangement to cover their Jakarta operations six months later.
" We'd been trying to hire for those three entry positions for months and couldn't fill them. Outsourcing removed the problem entirely and the team we now have is focused on customer service rather than typing. — Operations Director, Freight Forwarder
JATS XML Conversion — 220 Journal Articles for PubMed Central Submission
Client: Biomedical Journal Publisher, USA
A US biomedical journal publisher had a six-year back-issue archive of 220 articles that needed to be converted to JATS XML for PubMed Central submission as part of an open access compliance requirement. The articles existed as Word manuscripts and typeset PDFs. The journal had no in-house XML production capability and previous attempts using automated conversion tools had produced XML that failed PMC validation.
We confirmed the journal's JATS version requirement and PMC submission specifications before production. Each article was converted from the Word manuscript source with full element-level JATS tagging — front matter (journal metadata, article IDs, author affiliations, funding statements), body structure (section hierarchy, tables, figures, equations), and back matter (reference lists, acknowledgements). References were tagged at the element level with DOI and PubMed ID fields included. All output was schema-validated before delivery. A PMC pre-submission check was run on the first batch and the one schema issue identified was corrected before the remaining batches were processed.
All 220 articles were delivered as PMC-compliant JATS XML packages in eight weeks. The initial PMC submission achieved a 97% pass rate on first submission — the three articles that required revision had missing DOI data that the publisher confirmed was not available in the source manuscript. The publisher's compliance deadline was met with two weeks to spare.
" We'd spent three months trying automated tools and getting failed submissions. Getting to a 97% pass rate on first PMC submission was better than we hoped for. — Editorial Production Manager
CV Formatting and Candidate Database Build — 1,800 CVs Reformatted for ATS Migration
Client: Executive Search Firm, UK
A UK executive search firm was migrating to a new ATS. The existing candidate database contained 1,800 CVs accumulated over seven years in 14 different formats — some in branded agency template, some in raw candidate format, some as plain text emails and some as LinkedIn PDF exports. The new ATS required CVs in the firm's current branded template in a consistent format for correct parsing. The migration window was four weeks.
We established a conversion workflow with the firm's template as the output standard. Each CV was converted to the branded template with correct section ordering, heading styles, date format standardisation, employment history sequence confirmation (most recent first), qualification formatting and contact detail placement. CVs submitted as plain text or LinkedIn exports required the most manual reconstruction work and were processed at a lower daily volume with quality review before delivery. The firm's senior consultant reviewed a sample of 50 CVs from the first batch before full production continued.
1,800 CVs were reformatted and delivered within 16 business days — within the migration window. The ATS migration completed on schedule. The firm's consultants noted that having consistently formatted CVs improved their ability to compare candidates during shortlisting — something that had been difficult with 14 different original formats.
" The time saving was significant, but the unexpected benefit was how much easier shortlisting became. When every CV is in the same format your eye goes straight to the relevant information. — Managing Director, Executive Search
B2B Contact Database Build — 6,000 Verified Contacts Across Three Industry Verticals
Client: Market Research Agency, Australia
An Australian market research agency had won a contract requiring primary research across manufacturing, logistics and professional services companies with 50–500 employees in Australia and New Zealand. The contract required telephone interviews, but the agency had no existing contact database for the target population. Building the contact list internally would have taken the research team eight to ten weeks — longer than their project timeline allowed.
We researched and compiled contacts from LinkedIn, company websites, industry association member directories and professional registries — capturing company name, industry classification, employee count band, headquarters city and state, primary contact name, job title, seniority level, direct phone where publicly listed, email and LinkedIn URL. Each contact was verified as current at the time of research. Australian and New Zealand geographic coverage was confirmed for each industry vertical before production, and the client's target employee count bands were applied consistently across sources.
6,000 verified contacts were delivered across three weeks — 5,200 in Australia and 800 in New Zealand, split across the three verticals as agreed. The research team began telephone interviewing within four days of first contact delivery. The agency's project director confirmed contact accuracy (reachability on the listed phone or email) at 84% — above the 75% minimum the agency used for contract planning.
" We'd have lost the contract if we couldn't start fieldwork on time. SDES delivered a list that we could actually call — not just names with wrong phone numbers. — Project Director, Market Research Agency
Have a similar data challenge? Start with a free pilot.
Send us a sample of your source files and describe the output you need. We process a free pilot batch and return the output so you can review quality, accuracy and format before any paid production begins.
- Your own sample files reviewed by our team
- Output prepared in your required structure and format
- Quality, accuracy and workflow assessed before commitment
No commitment. No contract. Pilot returned within 24–48 hours.