OP-014
An established catalogue can need substantial improvement without everything in it being wrong.
A product range may have accurate specifications while similar products are still difficult for Shoppers to compare. Important attributes may exist but appear under different names. Some descriptions may contain the information Shoppers need, while others leave obvious gaps.
Images may be strong. Product identities may be dependable. Categories may still make sense.
Yet enough weaknesses can accumulate that Shoppers have to do more of the work of finding, understanding, comparing or choosing products.
At that point, starting again may look attractive to the Operator.
Instead of working through years of different treatments, they can imagine rebuilding the range around one new structure, one current standard and one consistent way of presenting the products.
Sometimes rebuilding the range may be justified.
But the fact that a catalogue needs substantial improvement does not establish that everything already there needs to be replaced.
It creates a different question.
When an established catalogue needs substantial improvement, what should determine what the Operator keeps, changes or rebuilds?
Consider a range of office chairs.
The product pages contain accurate dimensions, materials, warranty information and useful images.
But the products are difficult for Shoppers to compare across the range.
One product page labels maximum user weight as Weight Capacity. Another labels it Maximum Load. Several include the information only in the description.
Seat-height ranges are represented differently, and adjustability features are described inconsistently.
The range needs improvement because the Shopper has to do too much work to compare products that should be relatively straightforward to evaluate.
From that comparison problem, the Operator can see what needs to become better for the Shopper. They still do not know what should be rebuilt.
Replacing the product images would not resolve the comparison problem. Rewriting accurate warranty information would not make the relevant specifications easier to compare. Discarding dependable dimensions simply because other attributes need substantial correction would mean changing something that was not part of the problem.
Here, the improvement is the better outcome for the Shopper.
Before deciding what to keep, change or rebuild, the Operator needs to identify which catalogue elements are contributing to the comparison problem and which of those they can still rely on.
Only then can they decide how far the catalogue needs to change.
Knowing that improvement is needed does not, by itself, tell the Operator what to change.
Return to the office-chair range.
Suppose the dimensions came from current manufacturer specifications and have been checked against the products. The images are current. The model names are distinct and correct. The warranty information is consistent with the supplier documentation.
The Operator may already be able to rely on those parts of the catalogue.
Keeping them does not mean accepting the catalogue as it is.
It means recognising that the problem being corrected does not extend equally to everything the catalogue contains.
The problem remains in the inconsistent comparison attributes.
If seat height, weight capacity and adjustability matter to the Shopper’s choice, the Operator needs to represent those values clearly and consistently across the range.
Shoppers can then compare the products more easily.
The optimisation may therefore involve substantial changes across the range while leaving other parts intact.
A large improvement can still be achieved without making every existing catalogue element new.
But preserving existing material is not automatically the better decision either.
Suppose the Operator checks the weight capacities in the same range.
Some values came from old supplier spreadsheets, others from discontinued manufacturer pages, and others from descriptions written several years ago.
Several values conflict.
The information already exists, but the Operator may still not be able to rely on it.
Standardising the label would make the range look more consistent without resolving whether the values themselves can be relied upon.
Some values may be confirmed from current manufacturer information. Others may need clarification or replacement.
The same reliability test applies to structures and relationships elsewhere in the catalogue.
The products may still be grouped appropriately within a useful category. A compatibility structure containing unresolved conflicts may not be reliable.
An attribute may still represent a meaningful product distinction. Another may combine several different meanings under the same field and require restructuring.
Whether something is old or new is therefore not the useful test.
What matters is whether the Operator can still rely on that information, structure or treatment in the improved catalogue.
Now consider a replacement-parts range.
The Operator finds duplicate product identities, conflicting compatibility information and several classification conventions introduced by different supplier feeds.
The same part can appear under different names. Some compatibility relationships cannot be traced to a dependable source. Products that should belong together have been classified differently, while apparently similar products cannot safely be assumed to fit the same equipment.
Local corrections may improve individual pages.
They may not resolve the underlying problem.
If the Operator cannot establish which compatibility relationships or classifications can be trusted, local repairs may not be enough.
Correcting those compatibility relationships or classifications one at a time may leave the same underlying uncertainty in place.
Rebuilding the affected relationships or classification structure may then be the more appropriate form of optimisation.
That still does not necessarily mean discarding the whole catalogue.
Accurate product identifiers may remain. Current images may remain. Confirmed technical specifications may remain.
The unreliable layer can be reconstructed while dependable material remains part of the improved catalogue.
A rebuild can therefore be part of optimisation rather than evidence that optimisation has failed.
The Operator should decide how far to change the catalogue based on what they have established about its condition, not on a general preference to preserve what exists or replace it.
Knowing what needs to improve for the Shopper does not, by itself, tell the Operator how much of the catalogue needs to change.
An established catalogue can contain dependable information, structures and treatments alongside others that are no longer reliable.
How much the Operator changes therefore depends on what they can still rely on.
Catalogue elements that remain dependable can stay.
Those that need correction can be changed.
Elements the Operator can no longer rely on may need restructuring or rebuilding.
A catalogue can therefore need substantial improvement without requiring the same degree of change everywhere.
Suppose an Operator reviews a category of twenty dining tables.
Shoppers can see the products clearly, but comparing them is unnecessarily difficult.
Dimensions use several formats. Seating capacity is missing from some products. Extension mechanisms are described inconsistently.
The timber species and finish information, however, is current and dependable across the range.
The Operator does not need to begin by deciding whether to keep the existing category or rebuild it.
They can begin with the improvement required:
Shoppers need a more dependable way to understand and compare the tables.
The Operator can then identify the catalogue elements involved in that comparison and check which of them can still be relied on.
The timber and finish information can remain.
Dimension values can be standardised where the underlying measurements are confirmed.
Missing seating capacities can be established where a dependable source exists.
Different descriptions of extension mechanisms can be reconciled around a consistent treatment.
If the category structure itself prevents useful comparison, that structure can change too.
The result may involve hundreds of catalogue changes while still preserving a considerable amount of what was already there.
Instead of asking how much can be kept or how much should be rebuilt, the Operator can ask:
What needs to improve for the Shopper, and which parts of the catalogue can the Operator still rely on?
From there, the Operator can determine what should remain, what should change, what should be restructured and what genuinely needs rebuilding.
With an established catalogue, the Operator rarely faces a clean choice between keeping what exists and replacing it.
They can make that decision in sequence: identify what needs to improve for the Shopper, identify the catalogue elements involved, establish what can still be relied on, then decide how far the catalogue needs to change.
A significant problem does not automatically make every existing catalogue element a problem. Equally, the Operator should not keep something simply because it is already there.
That allows them to make selective changes without assuming the overall change should be small, or extensive changes without assuming everything should start again.
The appropriate response may be a small correction, hundreds or thousands of catalogue changes, or the reconstruction of an affected part from the beginning.
The Operator’s job is not to minimise the amount of catalogue change or to maximise it. It is to identify what is already sound, what needs to improve for the Shopper, and how much of the catalogue needs to change.