Cleaning up legacy data: Modernising your old product catalogue

Your legacy product catalogue is a bit like an attic: There’s some useful stuff, a number of familiar and beloved objects, but it’s also full of things you’re not entirely convinced you still need. And sooner or later, you’ll get round to having a good sort-out. Transpose that to product data. Today’s digital channels, marketplaces, and customer expectations aren’t working on the premise of information based on a historical scrapbook of half-matching spreadsheets. They want, in fact, demand, demand clean, consistent, richly structured data. And you may have some of it, but much of it is legacy data, and disorganised to boot. 

What to do about that clean-up? The good news: you don’t need to fix everything at once. Below, we outline how you can take the drudgery out of getting your product catalogue into better commercial shape by dealing efficiently with legacy data.

Why legacy catalogues block modern growth

Firstly, the majority of established businesses didn’t start with a modern Product Information Management (PIM) system. Their product data grew organically across ERPs, shared drives, local databases, and “temporary” files that somehow managed to survive various reorganisations.

That leaves those businesses with four problems which just won’t go away:

• Siloed data and conflicting versions: Pricing might be correct in one place, specs in another, imagery elsewhere.

• Inconsistent structures: Product attributes have been duplicated, renamed, and even captured as free text.

• Missing modern fields: What’s needed nowadays? Sustainability data, rich media, channel-specific requirements, and detailed compliance info. With legacy systems, part of or all of this information is often absent.

• Weak discoverability. If your data is in a mess, your internal search and site filters are also mired in chaos. As for SEO, well, cross your fingers and hope for the best.

In a sentence, it’s not just that legacy data is old, because it also causes the biggest obstacle on your journey towards better product experience and faster time-to-market.

So, what steps can we take to address these problems?

Step 1: Define the “desired” state of your catalogue before you touch the remnants of the past

The best way to eat away at a legacy clean-up budget is to get going on that process without having a clear idea of your target model.

Before you migrate anything, it’s crucial to define your future. Consider these areas:

• The structure of your product taxonomy and categories
• A standardised attribute model along with clarity of definitions
• Controlled vocabularies and unified standards for measurement units
• Mandatory data fields for your key channels
• What your “golden record” of all product should actually contain (as a minimum)

The above give you a blueprint for what data to keep, what is fixable, and what to retire with honours.

Step 2: Audit with intent, not panic

Don’t be in a hurry. Act mindfully. A legacy audit needs to be more forensic than frantic.

Map the locations where product information truly lives and identify the most reliable source for each data type. You can then profile the data for the following key metrics for quality:

• Completeness
• Duplication
• Inconsistencies in formatting
• Whether outdated
• Whether there are redundant fields
• Category relevance
• Revenue impact

Your primary goal is to prioritise sets of data. High-volume or high-margin ranges deserve first attention. Lower-value can get fixed further along the line.

Don’t get in a panic because it seems overwhelming. In fact, this is where a ‘Minimum Viable Product (MVP)’ mindset saves your sanity. Just remember you’re building an incrementally cleaner and stronger catalogue for optimised business outcomes, not chasing catalogue perfection yesterday, just for the sake of it.

Step 3: Clean, standardise, and de-duplicate in a safe staging area

The old “Lift and shift” approach is how old problems get a free pass. What should you be doing instead?

1. Extract from each legacy source
2. Transform in a staging environment
3. Load into your new platform only after validation

The key elements inside this process include:

• standardising units and formats
• stripping messy legacy HTML
• aligning free-text values to option-limited picklists
• merging duplicate SKUs into a single ‘best’ version
• splitting unfeasibly large fields (such as a giant-sized “features” cell) into clearly-structured attributes

Automation and AI can significantly accelerate this phase, especially when you’re dealing with large catalogues where manual cleansing would take infinite patience and lots of coffee.

Step 4: Enrich for modern channels

OK. Now you’ve got a clean structure, you can move from “accurate” data to “competitive.” That means putting in place what legacy catalogues often lack:

• high-quality imagery and video
• channel-ready copy
• sustainability and compliance data
• detailed technical attributes tagged for filtering and search

Modern PIMs have features enabling you to manage completeness scoring as well as flagging product data which falls below clearly-established quality thresholds. Carrying this out helps you tackle gaps systematically rather than randomly.

If your structured attributes provide a solid informational foundation, AI-assisted content tools can generate or refresh product titles, bullets, and descriptions. What’s more, they can do it at scale, alongside human oversight, and spot checks to keep the tone and any compliance and sustainability claims honest and verifiable.

Step 5: Lock in governance so history doesn’t repeat itself

We mentioned thresholds above, and that’s where governance enters the fray. A clean-up without governance is like deep-cleaning your house while a bonfire blows embers and ashes in through an open window. Your main objective here is to prevent regression. Do the following:

• Assign clear ownership by category
• embed validation at point of entry
• enforce controlled vocabularies
• schedule regular data health reviews
• Make sure that any integrations with ERP and eCommerce don’t reintroduce inconsistent values

This is the moment when your PIM becomes more than a system of record. Rather, it becomes a system of data discipline.

Where Start with Data can help

Legacy modernisation is rarely just a technical exercise. It’s a business transformation programme with a data backbone.

At Start with Data, our experts help your organisation to take a pragmatic route through this potential minefield:

• auditing and scoping legacy landscapes
• defining the future taxonomy and attribute model
• prioritising high-impact categories
• using AI where it makes sense to accelerate cleansing and enrichment
• implementing and integrating PIM to support long-term quality

The desired end state is always the same: a single, trusted, scalable product catalogue which will support your growth plans rather than slowing them down.

Final words

If your legacy catalogue is slowing launches, undermining search, or making PIM implementation feel more of a pain than it needs to be, Start with Data is here to support and partner you in replacing the disorder with order. We offer end-to-end services to support modernisation, from forensic data audits and taxonomy design to AI-accelerated cleansing and enrichment, plus implementation and managed services.

Get in touch with us today and let’s have a more detailed conversation about how we can help you build a phased, high-ROI plan which will rebuild the foundations of your product data without derailing day-to-day operations.