Transcripts

Transcripts are generated using a combination of speech recognition software and human transcribers and may contain errors. Please check the corresponding audio before quoting in print.

 

 

Welcome to Voices in Local Government, an ICMA podcast. My name is Joe Supervielle,

and today I'm with Rahman Shah, President of Rahman Shah Data Science and regular

contributor with ICMA, ELGL, and more. Thanks for joining. - Thanks so much,

Joe. - Today we're gonna get into specifics about local government's all too common

software trouble and how to fix or maybe even better a vo -

Tell us what is good data? Where is its place and sound decision making for

government be it the city county managers themselves or even their staff? How can we

move from that just being a cliche or like an effort that we kind of try to

something that's really happening? I think the slowdown That's kind of you know The

the heart of it, you know as a data scientists, about 80 % of every data project

is getting your arms around the data, right? Which is like, does what sitting in

bits and bytes on a hard drive actually, how does it correspond to day -to -day

operations? And then once you have clarity there, I mean, the war is almost over,

it's almost won. And I see people in this world, we've been a data just for about

11 years now, I've seen over and over, people kind of get, they get ahead of

themselves, right? You get impatient and want to plug into things that are very

sophisticated without necessarily kind of even being like, do I have a record of

reality that I can trust, right? And that's kind of where that's where all the

value is. And that's the thing that keeps not happening, and I think that's why

people across sectors honestly just keep failing to avail themselves of the

opportunities that the data can provide them. So what would be the first step to

slow down and get it right? I think it's you want to get some increment of making

an operation better with data. It does not have to be fancy. I think that's the

thing that's like, that's the thing I want to drive home. Often, especially as you

get to larger governments, right, you got into these, and this is, you know, segue

right into what we're going to talk about in the core of this conversation. You

know, you start putting a bunch of infrastructure around it. Now you've got this,

like, you know, grand design and like a big procurement. We want to be data driven

as a city, as a county. You know, we hear that that happens every day, right? But

can you actually close the loop, close the circuit all the way from like, I was

doing something this way, and then I measured something and looked at some data, and

then I changed how I'm doing it, and I can show that I've done it better. Once

you get a taste of that, you can kind of like, that becomes a little bit of a

flywheel. And then you can really-- - Yeah, incremental improvement. - Absolutely,

right? And I think that that

as educates a lot of stakeholders, a lot of leaders, a lot of frontline workers,

a lot of people who are not necessarily sort of technical or not like data

scientists or IT people. That starts to educate people about a lot of the gremlins

that exist. That the thing in my Excel spreadsheet, the thing in my Google sheet,

the thing in my IT system may not actually align with reality. And I need to clear

out all those gremlins, and that's not trivial. And I think getting to where you

can actually complete the thing as little and modest as possible gives people a

taste for that, right? That there are always so many gremlins in the data, and even

getting to where you have some kind of digital record of reality that has any kind

of bearing on something you can work with. That's the whole process. Okay,

so as you said, that's going to tie in to our main topic today, which is software

trouble. Yeah. Define it. Let's start there.

Governments procure, they buy, they have systems. Sometimes their staff need it.

Sometimes they want it. Oftentimes they don't. Or they say, "We needed this." Not

that. So we'll get into some specifics and how to avoid or improve it, but just

start off with your experience on what is software trouble and how do governments

kind of keep finding themselves in this position? Yeah, so how I define software

trouble is that you've got overly complex software deployments.

They have a really big scope and a lot of complexity. Overly is,

of course, objective, but then I think the second part is sort of what defines it

is that they end up costing too much, they end up taking too long to deliver.

Sometimes they don't launch at all. And then the depressing shadow of it is like

you have some fiscal trouble because you went way over budget, but then also you're

started to kind of hollow out your human capital. You've got staff that are burning

out and leaving, and it's honestly really sad, right? Because the government starts

to lose its institutional capacity over this kind of albatross of a software

deployment. That's the human cost of software trouble that I think is perhaps most

salient. Okay. So what is the starting point for finding yourself in that position

with bad software? And I think even before there's any purchase,

that idea stage or what do we actually need can maybe go off the rails sometimes.

So start us off from the beginning of a, you know, the hypothetical of what went

wrong, walk us through that so the listeners can avoid it. Absolutely. So the thing

about software trouble that I keep running into is that it's sort of root cause is

like before anyone started to pay attention. You've got some kind of amorphous set

of meetings, you've got some email thread, you've got something like that where

There's not a whole lot of structure around trying to define like what are we even

trying to accomplish? Like what is what is what could get better around here? and

so you have this kind of Self -assembled poorly controlled brainstorming process that

has sort of happened before before really anyone's paying attention and Through that

process people end up with this kind of overly expansive idea because it costs

nothing to throw something extra into the pot, right? It might make you look good

to your boss if you throw in something else. Or you, so you have this kind of

like, it's just kind of just humans being, you know, we're primates after all,

right? Just primates being primates, right? You know, put them in a room and they're

gonna do a thing. And this is the thing that happens. - Quickly, when we're talking

about the brainstorming and kind of Part of the idea there is no idea is a bad

idea people can throw stuff out see what sticks But how do you move from? Everyone

contributing and also I should point out when there's so many different stakeholders

or teams and then individuals within those teams Someone can have a legitimately good

or fair idea that makes sense for them But it doesn't necessarily mean it's a

priority. So absolutely. How do you how can a group? better organize idea stage,

shifting down maybe like a funnel. I'm visualizing a funnel back to what we actually

end up with to move forward. Yeah, so, you know, this is thankfully something that

a lot of smart people have thought about for a long time, right?

The process of adding structure to brainstorming and idea generation is like, you

know, it shows up across all of these different disciplines, right? We've got the

lean startup guys, right? West Coast tech guys, you know, and they talk about

minimum viable products, right? What is the tiniest, crappiest possible increment of

something that could possibly test if this thing even helps anybody, right? And then,

you know, now we've got the design thinking guys from a completely different

tradition, right? And they say, okay, you know, you do divergent thinking then you

do conversion thinking, right? And then that turns into a prototype you can rapidly

deliver. You've got the Lean Six Sigma folks, you know, mostly from the manufacturing

world. They sort of say, okay, brainstorm, but then, you know, we're going to use

these things like affinity diagrams, multi voting, we're going to use that to bring

focus to an initiative. So I'm not generally married or a huge fan of being like a

super in one camp of branded methodology, I think they can turn into cargo cults

and not accomplish much. But like you've sort of seen that thoughtful people who've

tried to make things go well, all converged on this idea that by giving structure,

you have there be even less penalty for having bad ideas and brainstorming, right?

Like let's get it all on paper, but then let's remove the shame, remove the status

from, from now, like, okay, this is a good idea, but it's out of scope for now.

We'll do this a different time, a different place, a different way, so that you can

really kind of come down onto a great deal of focus about what you can change,

right? And it's very easy when I talk to people, like, what's, what's, what's,

what's hurting you in technology? And they're just like, everything, we just need to

blow it all up and replace it all. You know, but then I poke at that. And then

eventually they're like, our financial reporting is like really difficult.

Okay, now that's something we can work on, right? As opposed to everything, or oh,

I want to just blow it all up and start again. I want to do this sort of

wholesale ERP migration. ERP is enterprise resource planning and is one of the kind

of key sources of software trouble. - So you are suggesting a scalpel approach as

opposed to the hammer, what have that metaphor supposed to be? Maybe, yeah. I think

it's an approach of having to get a good inventory of where are there problems and

then trying to really come up with the most concise possible description of what

that problem is and how you might be able to solve it. I would suggest too is I

have a heuristic for how do you, because you might end up with Several possible

ways you can go and I think if you can get someone who has a you know where you

have a really passionate internal champion For that piece of change that's gonna make

a huge difference and that's usually a person who's like at the bottom of your

oratory They're very close to ops And they've been around for a while. They're gonna

stay around for a while So not gonna like quit like you know six months into a

significant piece of work and they will I've almost always seen that like they will,

they have knowledge that leadership needs to know about. Because they're still in it

day to day. They're using. Yeah, absolutely. Right. And they're making them better at

their job is the ultimate goal of the thing. And they often are not, not consulted

until, you know, until, okay, they've had something foisted on them, it doesn't work,

right? It's made their, you their their their productivity plummet. And so now

they're trying to kind of scream their way through like three layers of

counterparties, right? Not only people wave toward the top of their hierarchy,

but then maybe a software vendor and maybe a system implementer. And they're trying

to expose facto clean up the mess when they don't really have much leverage. But if

you can really focus an initiative, brainstorm yes, but then window and focus, and

identify a specific thing you can do with a passionate internal champion. You're

gonna have a lot less trouble and probably win the war before any of the really

big trouble comes in. Okay. And that whole process of distilling it down to the

primary need and then how to get there, that is how you avoid what I referenced

earlier about just letting it go off the rails where it just gets it and then

mission creep and then different perspectives get jumbled in and there's internal

politics and it's not for trouble as the episode is titled. So avoiding this through

distilling down the ideas into a clear and obtainable goal,

is that fair for maybe kind of step one? I think that's a good way to put it.

Okay, well let's say that was not, that did not happen. We're kind of stage two

already, and it might be a little messier. It might be a little bloated. But yeah,

why is the purchase or procurement process so complex and perhaps ripe for mistakes?

Yeah, you know, so this is like a big and weighty problem, right? Which is not,

doesn't have any use. It's just like, we have this collision of like, you know, the

public service motive motive of the local government manager, you know,

colliding into the profit motive of capitalism, right, through a procurement

transaction. That's a tough ballwax, right. Navigating that tension is not easy.

Where a local government leader has a ton of leverage, right, is that it doesn't

have to be this way. Like there are a lot of things that capitalism does for us

phenomenally well. The thing you do is you start to talk to representatives of

vendors and start to kind of co -create some kind of solution with them is you have

to realize what are some of the things where transactional capitalism is good and

where are some of the places where it's not so good, right? So there's a lot of

profit potential in sort of representing something that is a big mashup where maybe

a corporation bought a bunch of little startups and just kind of clued them all

together. They're not really integrated and just representing them as if they are.

And saying, oh, like there's all this stuff that's causing you stress, right? Like

systems are complicated, processes are complicated. They are a messy mashup under the

hood. And we're gonna just like, you know, kind of pristinely replace it with this

Swiss watch. But it's kind of fake Swiss watch. So you wanna look for the real

Swiss watch, right? Which are things that are, again, very simple. Like whenever I,

you know, ask, you know, I ask local government leaders, like where, you know,

where in your software systems, you know, do things like kind of amazing delight

you. Always come back with things that are like very, very narrow in their,

in their ambitions, right? Like, oh, we have an application for, you know, parking

enforcement. And it's just great. It's so much better than what was or set it,

forget it, we trust it, we're good. And things like that, I think,

is where you start to see a good future for kind of like technology in local

government, right? As opposed to like, oh, you know, we kind of got the Walmart of

digital stuff, right? And it's gonna kind of do everything. Things start to get so

complex that nobody can really understand how they work. Not even the software

engineers at the vendor understand how they work. I know that as a former software

engineer of a vendor. - So are you describing kind of the difference between a

platform versus a tool? Is that fair? And if so, are you suggesting the tool is

more effective than the overarching platform? - Yeah, so gosh, I'm so allergic to the

term platform. It's just, you know, my alarm bells all go off when I hear hear the

word platform. And yes, like a tool is a thing that you can understand,

that you use to do a thing.

Tools are amenable to people being able to describe what they do and how they could

be better. That creates a nice feedback loop between public servants and then the

vendors who serve them so that you and then improve that tool and make it better

and better. Platforms tend to instead be kind of like, oh, well, this is a big

expensive thing. It's kind of optimized to be, to like milk a lot out of a

procurement, right? Milk a lot out of a customer relationship. I hate platforms and

I like tools.

- Well, fair and honest answer, good. So how do you, you said push back about how

can the local government, Not necessarily representative, but the person kind of

leading this task What is their leverage? Mm -hmm, and how can they maybe find some

common ground? Regardless of the size or scope of the vendors, but how can they?

Better avoid overdoing it. Yeah You also have a lot of vendors who are solving,

you know a specific problem, and I think pricing yourself toward them is a priori,

a good idea. You might actually be acquiring a platform. Maybe it's not in your

call, right? I mean, I'm sure many of the good listeners of this podcast are not

like, don't have top level responsibility for a huge procurement, but are somewhere

else in the mix. And like, if you've got someone who's offering, you know, a big

complicated platform, In fact, maybe you got pulled into that anyway because you

bought one of these specific tools and it got bought by a big company. Don't feel

forced to use everything they offer. You can pick and choose a la carte within your

thing. Some of the pieces are probably awesome and then some of them are okay and

some of them are terrible. You can definitely stitch together something that works

for you that only uses part of what you bought. Of course, it would be great to

know that before time and make sure that what you're paying is commensurate with the

value you're hoping to enjoy from it. But a software system is not a religion,

you know, it's not this kind of, you don't have to, you don't have to buy into

all of it to be using it. And you can often benefit from a piece of it. And

that's really useful. And then, you know, third, you know, this is,

I think think some such a huge value of professional society is that, you know,

your peers are going to know things. They might have used something and have a take

on how it went. And like, they don't have any incentive to, to misrepresent that to

you, right? And so asking people what software do you use, how do you like it is

helpful. And that's, you know, sort of also, that's synergistic with the bias towards

small, simple tools, right? If you're looking for a parking enforcement application,

that's the kind of pointed question that can give you a real sense of the

competitive environment. While it's like, oh, we're just kind of digitizing all the

things, that's gonna look different for every jurisdiction. And there's gonna be a

lot of customization, there's gonna be a lot of

you know, that's done one off, one off. And that's gonna make your vendors a lot

less, it's gonna make your peers' experiences a lot less comparable. - Right, okay.

So stage one was all about moving from ideas that might be all over the place to

a coherent and focused, this is what we need and why and what we hope to get out

of it. - Exactly. - Step two, whether step one went well or not.

do we actually need out of this thing? - Yeah. - And how do we use it as a tool

rather than a magical tool? - And choose vendors as well. - Probably isn't. - Right.

And like, you know, choose people who are gonna really work on your behalf and

great, you know, through peer advice and other tools, right? - Yep. - Okay. At stage

three, so let's pretend that stage one and two still did not go well and maybe

some of the listeners find themselves in this right now. - Yeah. - You've found, you

know, the idea ideas kind of went sideways and then the procurement ended up with

this big, fancy platform that gave you that reaction, that negative reaction. Big

promise, really competitive bid, right? Yeah. So stage three is the, this is not

going so well, but that human tendency to double down. Robin's got some articles

coming out that will be linked wherever you're listening, a little series on, and if

you if you want to keep learning about this but in it you include a quote from

author and respected thinker Kurt Vonnegut and that goes a step backward after making

a wrong turn is a step in the right direction. Yep. Which we can all understand. I

will add in a quote from the famed philosopher Lloyd Christmas who once said we're

in a hole we're just gonna have to dig ourselves out.

So So sunk cost fallacy, I think we all kind of understand this intellectually, but

it's different when it's real, fear, pride, ego, reputation, avoiding either potential

or real perceived chaos. The money, the budget, the public's money is at stake. So

again, if the listeners find themselves in this position of, okay, we overdid it,

this is not going well, the contract is the contract. We're kind of coming to the

decision point on, do we really start over? Or do we just try and make the best

out of what we have? Yeah. Walk us through what you've seen in your experiences

advising local governments on just this key point or this tipping point. You know,

Joe, this was the thing that I think, it's just heartbreaking to see. And this is

a thing that kind of got me into a place where I spent a lot of time thinking

and advising local governments on software trouble. Like, I'm a math guy,

right? I got into this business to do math, right? I'm like a PhD scientist by

training. And yet, you know, it's in this process of doubling down, right?

There's, you know, mistakes were made that were maybe not even reckoned with at all

at this point. And then things are starting to kind of be, oh, what we were

promised is not really, you know, is not really what's happening, right? And it's

this process that this is really where you start to just like really churn the guts

of an institution, right? Is when you start having to face this thing that it's

like, someone has staked their reputation on this process working, right? They've been

forced to make commitments outside of themselves to, okay, we're going with this

brand, we're doing this thing this way. And now it's not working. because you were

maybe, you know, a lot of that might be just kind of good faith issues that were,

you know, came up. And then sometimes there's a degree of misrepresentation in the

sales process. And so it's like, well, you know, now it's like, how do you, how do

you lean in into your teams, as opposed to leaning against them, right?

When you're under threat. That's really hard, right? You know, I've been in this

long enough to know the kind of pressure that like a city manager or county

administrators under, right? They've got a lot of pressure from the public, from

elected officials, but then they've also got to sort of hold this team together. And

then amidst all of that, they have like kind of, if they have kind of an

excessively cozy relationship with a vendor who made a bunch of sweet promises to

them, they've we've been coaxed into a place where they're sitting on the same side

as the vendor, right, against people like their teams. And often those low -level

front -line workers who are the internal champions whose job they're actually trying

to protect. And there's a real danger there. And this is, of course, this is done

on purpose. I mean, a successful software sales rep makes,

frankly, a lot more than They make a lot more than you. They make a lot more than

any city manager I've met. You know, it's a very lucrative business. And there's a

great deal of expertise going into this delicate psychological game of getting people

to commit. And then when things don't turn out, getting them to instead of being

like, there's something wrong, let's stop the bus. Being like, oh, let's dig a

little deeper and keep going so everything looks okay. Let's make it-- - Keep

digging, yeah. deep digging, right? And so, you know, the, you know,

and honestly, it, you know, from a personal level, it's, it was very, this, this,

this was very hard to unsee and very influential. Like, you know, I, early in my

path as a data scientist, I was actually a software engineer, right? That's how I

started. I was coding full time. I was making web applications that did data

science. Equations of mine was like, you know, always leave a hook in your work.

Yeah, we should always make sure something doesn't work quite right because if

everything works perfectly, they'll never call you back Hmm, which is very much the

opposite of Swiss watch thinking, right? And like I realized that there's this kind

of like element of bad faith that is like very prevalent unfortunately in the

software business and

You know, I was just like at some point. I was like forget this. I'm gonna do

something else You know, that's kind of where I started leaning toward this kind of

continuous improvement, process improvement work that is really where I sort of,

that's the perspective from which I do data science now is transaction, you know,

transactional capitalism, you know, encourages people to weaponize sunk costs.

It's always better to promise more and quote less and then just have them kind of

like take a bath on the back end than to, You know make a honest promise and an

honest quote and just not get chosen because you're not you're because you're not a

competitive bit That is you know, that's that's a really big problem and it was

something that I it's so it's so damaging And if I could do anything to help

people sidestep that I I will

So how can the leaders whether it's the city county manager or department head or

just a staff person running point on some of these implementations. Yeah.

When you talk to them, is there a light bulb moment where they can back off and

kind of take a beat, take a second and remove the bias or some of those emotional,

psychological, I hate to say trappings, but to get out of that sticky part and say,

okay, this isn't working, not necessarily automatically, let's kill it all and start

over, but what do they need to hear or what do they need to do or what kind of

support or reassurances do they maybe need, whether it's from the council or

otherwise, whoever they are kind of reporting to, to actually make the right choice

and make a logical, going back to data -driven, the cliche, make a data logical

choice instead of an emotional one. How do they get to that point? - Yeah, so

getting to that point is something that does not happen in a day. And I would say,

thankfully or not thankfully, depending on how things have been going, it builds on

some of these much more ethereal aspects of leadership, of leading a group of

people. And so I would say a couple of those things. One is, if you have like a

really good articulation of institutional goals, here's what we're trying to

accomplish. And, you know, that's something that's been socialized well with your

team. This is usually what the job of strategic planning is. And I hope in many

cases that this is done successfully. You can, you can step back and be like,

Hey, like, okay, we've got this thing, we like, you know, we Aged brand X and sort

of system integrator Y to do this big messy thing, and it's not going so hot,

like it's a lot easier if you've been doing that groundwork of aligning an

institution around goals to kind of be like, you know, is this serving our goals?

If yes, proceed. But you might find that a lot of things are actually not serving

the goals, but have ended up becoming kind of being done for their own sake because

of all the steps it took to get there. And so having a clarity about what your

goals are, is it serving residents well? Is it saving labor at City Hall because

the general fund is tight? You can start to make kind of like educated decisions

about what to do. Do we proceed, do we edit the scope back,

right? Do we completely cut bait and just have to take a big painful, a big

painful L in public? Sometimes all of those things are true, but I would say that

people in, you know, software, software trouble is all about doubling down. So when

it's like you proceed, let's put more in, oh, it's over budget, over schedule,

everyone hates it. Okay, let's keep going. That death march is really where all the

damage happens. That's where governments lose their best people.

Sometimes not physically, but when they start quiet quitting, when they burn out,

when they become disengaged, that has a real effect on what a government can do.

And that all happens in that process of leadership doubling down to the detriment of

their teams. And that's really, I'd love you know, to any extent, I could stop

that, gosh, I wish I wish I could, you know, because it is so sad to watch. I

mean, the dollar values are horrifying for one 300, 400, 500, 600 % over budget,

right? Schedules are usually proportional to that. And those things are bad. And they

certainly, you know, we're recording this in very 2025, it has not been a pretty

couple years fiscally. And, you know, that's, that's one side of it, but human

beings are far less replaceable than we would like to believe, right? And there's a

very nauseating feeling when a dedicated public servant of maybe a couple of decades

is like, finally, just like I'm done and they walk.

That person's a lot less replaceable than we would like to think and protecting them

and making them feel safe and making them feel Like they're part of something

worthwhile is more important right then sticking with BrandX for some gigantic

software migration, and I think Keeping that in mind so that you can sort of sort

of step back and Regroup is really important and that's kind of the biggest antidote

I can think of to doubling down and and and causing all this all this trauma.

Okay. So again, this whole this whole episode, we've kind of given the hypotheticals

on when it goes wrong, but what you can do to solve it. And yeah, when it goes

right. So just in on a positive note, those instances where you have seen,

whether it's a single person or kind of a team come together and make that tough

decision. Yeah. How do they go about it in a good way? How can you mid failure

with grace and a plan to move forward. Have you seen, and not in a clean way,

it's not always going to be perfect. And there's going to be some collateral damage

probably, but how do you do that in the right way, including communicating it to

staff and potentially the public on, as you said, taking the L, but then explaining

what happened, why, and then what you're going to do differently, trying to get to

that point where the person or the group of people you're communicating to

understands this big L is still better to take now than to double it.

So how have people done this in the right way? Yeah. In the articles I've got

coming out, I interview one of my long -time contacts and friends,

Jessica Hoffman. She's an assistant city manager Well, she she went through this and

I got some quotes from her, but it's really kind of you have to message around

those goals. Like we did a thing, we took a risk and it is not serving our goals.

So we're going to make a change that does serve our goals. I think that's the

right way to do it. And that's why I think it's always so important to keep

communicating about goals, right? It's not about brand X, you know, and Brand X can

become a cargo cult. It can become a boogeyman, right? And in this conflict that

starts to transpire in this doubling down process. But if you can sort of keep

reinforcing that, like, we're here to do a thing. We're here to serve a public in

all of these many and highly cost -efficient ways that local government does, right?

Which is what dropped and drew me to the sector. You see a software initiative and,

You know, despite your best efforts and your best information at the time, it's not

working out or it needs to be edited. Maybe needs to do less or maybe needs to do

something different or needs to be just adjusted in some way.

Framing that in terms of what you're trying to accomplish, I think is so important.

And yeah, Jessica did a lovely job of that. She was able to, despite struggling

with software trouble, you know, she kind of learned about, you know, while I was

learning about it, honestly, right, you know, historically.

Yeah. Rahman Shah, thanks for your expertise with data software and how local

government can continually improve with each. For more on this topic, as referenced,

check out Rahman's blog series, Linked Wherever You're Listening, and can be found on

icma .org. And for those listeners who want to kind of keep learning about this or

just have some time with a real expert, you can connect and reach out to Rahman on

LinkedIn, also Linked Where You're Listening. So thanks again for your Joe, it was

such a pleasure.

 

 

 

In this episode of Voices in Local Government, Joe Supervielle speaks with data scientist, Raman Shah. Shah explains what data is (or isn't) and what to do with it. The conversation then moves to how local governments can best avoid software trouble and save millions.
 

Key Takeaways for local government's data and software:

• Be intentional during idea stage with disciplined winnowing.
• Understand the difference between a platform and a tool - and what will best serve your needs.
• Spot the sunk cost fallacy and resist the urge to double down out of pride, fear, or ego.

Featured Guest:

Raman Shaw – Owner, Raman Shah Data Science

Connect with Raman on LinkedIn
 

Voices in Local Government Podcast Hosts

Joe Supervielle and Angelica Wedell

Resources

Learn more from Raman in his three-part blog series, Software Trouble.

ICMA Annual Conference, October 25-29 in Tampa.