In addition, the strategy you've outlined is the one nearly every incumbent in every enterprise space has leaned on to protect themselves from disruption, but it only works for so long (as we've seen time and time again). I also think people overestimate the ability of large companies to quickly adapt their offering—the machinery of enormous sales, marketing, and product organizations can't turn on a dime.
In my experience cities have just as many procurement problems as state/federal but with far less money to spend. You may have success with a Dropbox/Hootsuite approach of selling to non-traditional users under certain price limits. It just makes it so much harder when it's your only market.
As to plentiful discretionary funds... everyone's definition of this is different. Another challenge I've faced with local governments is that their speed to spend is similar to state & feds. Smaller sales with a sales velocity the same as if not slower than the feds is another challenge.
But if you've got people to make fast decisions and spend enough so that the unit economics makes sense you have my heartiest congratulations.
But still it is no accident that ncd's analysis works at the local government level: there are tens of thousands of those governments so variation means there will be some viable customers around.
If you have a convoluted and slow purchasing process that favors existing vendors you are going to have problems. If you have that type of process and can't spend very much money up front, you are going to have worse problems.
Government, at all levels, has the second problem. This selects for large vendors (they can handle cash flow irregularities, have many low cost staff, have existing relationships and are on VoR lists/can easily pass credit/reference checks) and low quality. Price is especially key - bad software can be superficially cheap but "unexpectedly" need lots of consulting work or change requests. New high quality software is typically delivered by a firm that can't take advantage of post hoc consulting and change request revenue.
One easy way to block a startup with dropbox like pricing and market approach - privacy/security requirements. Make a state/federal law or regulation requiring all software touching citizen personal data be certified as safe. To ensure less waste and data security all software of any price must be provided by a firm on the VoR list. Purely for privacy and integrity purposes, of course!
When a government agency with minimal budget is the entire market, you face incredible hurdles to succeed that have nothing to do with product market fit. There's a reason why Uber and AirBnB broke laws & regulations and why government clients outside of defence & intelligence are typically the last targeted by a generalist software firm.