Model Size Tips & Other Modeling Tips & Tricks

As mentioned, unless you work in infra, you need to get comfortable utilizing more simplistic operating builds, especially when doing transaction underwriting. This is, of course, assuming that you are "owning" the model and that some VP or your firm isn't forcing these complete nonsensical pieces of shit to underwrite a midwestern services business (or a vertical SaaS business, a widget factory, etc...). 

A few random thoughts:

• Learn to be comfortable with hardcodes, and source your hardcodes for your own reference. If you're at any good shop with a lean team, your VP will assume your numbers are right and are not going to waste their time to audit every single fucking cell (unless you literally just started). I'm assuming this contributes to why you have so many fucking tabs for XLookups or external links pointing to original data sources. Just hardcode the shit, store the backup in the drive, comment, and move on. Especially if it's creating excel issues
• Be thoughtful and communicative upfront about how you segment and drive your revenues. For example, I've had discussions with my VP about product groupings when that asshole wanted to model out every single fucking SKU when we were in round 1 and had no clue how to make an informed view about it. Group products that tie into eventual commercial market diligence and that roll up into reported business segments. Collapse line items that make sense, and model them together. This greatly reduces complexity overall, reduces chances for errors, while still hitting the 80/20 view. 
• If the dataset is static, you don't have to sumif / lookup entire rows. OK if only a few, but if you do this every time then it takes some juice. Also, not every single thing needs to be a waterfall. Many times it does, but sometimes it's just way too much.
• [Maybe controversial, but my personal preference] Personally, unless it's an infra or infra-adjacent model, I'd avoid monthly builds for. Just use general conventions to estimate quarterly impact of monthly billings/bookings.  That way you don't need to roll everything up a million times, and let's be honest, even post-close the finance team isn't gonna be able to report all your little detailed metrics and financials every month, at best it'll be the biggest topline KPIs which you can track progress on the quarter
• Detailed headcount builds are overrated IMO. Group things where they make sense - you don't need 6 tiers of sales reps with different levels of managers, for example. Do what makes more intuitive sense. Sometimes you build to a target revenue number (triangulating with market data), and then you burden SG&A appropriately. Sometimes you start with SG&A / Sales Reps and you drive it bottoms-up. But include details on the fundamental core drivers, and group things that are finger-in-the-air estimates and drive those off %s.

This is just the start of reframing how to model in an 80/20 fashion. Every time I get a sell-side model that's 20+ tabs, it's almost entirely because they model out 20+ product lines that can be collapsed into 5, or they model every single fucking city of operation when they could do regional/market level builds. And every time I use the model as a reference, rebuild my own, and throw theirs in the trash because it takes 10 seconds to run a single assumptions change, let alone sensitivities.

IMO, if you need to wait more than 2 seconds for assumptions changes to run through your excel, you have too many tabs. Similarly, if your sensitivities are taking more than a few seconds to load, simplify that shit PLEASE.