Use your data source filter effectively. If your data set goes back to 2014 but you only use the last 6 months, set the data source filter up to only pull in ('extract') the last six months to cut down on what it's processing through the calculations.
Also use context filters after data source filters. This will also cut down on what the calculations are processing.
Maybe others have some more detailed ideas. But I'd start there. If you have a sample file to play with, I'd be more than happy to try and help.
Do you really need to place 27 measures together in one crosstab report? Tableau is an analytics tool rather than a reporting tool. You could make a measure name filter to let user pick what to be analyzed.
I presume by "54 placeholders" that you're using a multiple axis crosstab where you're using MIN(1) or AVG(0) or something like that to generate 54 axes. This horribly impacts performance in Tableau.
My suggestions are:
1) Do something like what Jian suggested and let the users pick from a smaller measure set, the less Tableau has to draw the faster it will go.
2) Go even further and take some time to figure out what users are actually doing with the data and build views that serve that need. Usually a giant conditionally formatted crosstab is something that comes out of Excel and is suited to what Excel can do, it's rarely a view that really suits the tasks that users are actually doing.
3) If you really have to then either pivot the source so instead of 27 measures you have 27x the number of rows, with a dimension then you can start doing more custom formatting. An alternative is to create a "scaffold" source that has a dimension for the measures as well as other required dimensions for display and potentially filtering and then use that as a primary source to a Tableau data blend with the original data as the secondary, and calculations that plug in the right values into the right cells. For example I once created a complex layout where I had a 1K row data source as the primary and the secondary was a 2M row source.
yes i have the use case of using 9 measures for prior week,current week and for the difference(as change) between these two weeks.
so i need to represent each of the category with different colors fillings on their headers.this is why i am using these many number(54 placeholders for overall 27 columns) of place holders.i had already used the context filters and hidden the unused fields.maximum tried to improve the performance but failed to achieve this.can anyone suggest me a solution for this?