I think what was meant was that you can utilize the cost approach as the basis to derive many different adjustments. It is a very useful tool to use when determining highest and best use, site value (using abstraction), effective age calculations, if a property is overbuilt relative to other properties in a neighborhood, plus many others. Oftentimes when other types of data are scarce, the cost approach can point out market changes and reactions that can be meaningful in making sense of the bigger economic picture in a particular market. Whenever possible, I develop the cost approach for my subject and all the comparables I've selected in order to test the results of other methods and adjustments that I've derived. It is just one of the many different tools that I use to perform as complete an analysis as possible on a property. Sometimes I use a simplified cost analysis tool (derived from the Craftsman Building Cost Calculator) to perform a quick cost estimate, while other times I might have to use the more in-depth Marshall Valuation System which takes more time to develop. It all depends on the property in question and the data available.
I think that there is this impression that the cost approach is inaccurate because the two most widely used sources of data don't accurately represent what actual costs are in a particular market. I don't dispute this, but even this can be overcome if you know how to localize your cost data to your market. I'm a data nut, so every six months I calibrate my cost data by observing and then adjusting for differences in the markets where I work. I do this by analyzing sales in new tract subdivisions where houses are typically built at the same quality level with mostly the same features. I use the un-adjusted cost data to estimate the cost of as many different sales as I can find in the neighborhood. From there, it is pretty easy to pull land to building ratios, calculate developers profit, and even spot potential paired sales elements. Finally, I calculate a local multiplier to apply to both the Craftsman and M&S data. It is surprising how accurate you can get once you localize your data. We also use this technique in mass appraisal all the time.
I know that for most appraisers, it's the amount of time that is required to properly develop the cost approach which is the deciding factor in its use (or non-use). The cost approach does take time to develop however it can be reduced considerably by making your own tools to streamline the process. I can perform a cost analysis for my subject and three comparables in about 10 minutes by using a spreadsheet that I made which replicates the methodology used in the Craftsman Building Cost Calculator. It's simple, easy, and fast. The real time savings however, comes from all of the other adjustments that I can automatically derive without any additional work (or added time). The effective age calculations shown in the images I've uploaded were performed automatically and didn't require any additional time since they were linked to the results generated by the cost estimator. This is only one potential adjustment which I now have support for in my workfile. There are many others. Thanks to the ability to link specific cells in my custom spreadsheets to input cells in WinTotal, I don't even have to reenter data, which saves me even more time,
To sum it all up, the cost approach might not be applicable to every assignment, but it can be an accurate and useful analysis tool that you would be foolish to discount completely out of hand without even considering its use.