Wednesday, August 28, 2019

Platform and Partners, Round Two

After reviewing this post on platforms and partnerships, there’s more to dig into. By definition, you can’t cross the Bill Gates line by yourself, but who should you be seeking partnership with? Developers who consult or consultants who develop? What tools should you build for them?

At the end of that article, I felt that free form coding was required. My reasoning is that the platform vendor cannot predict valuable use cases well enough to produce the toolkit that a consultant would need. This is not condemnation of the toolkit or the consultant. Rather, it is a recognition that high value jobs require deep linkage to the customer’s processes and data systems, meaning that they are complex and customer specific. This means you’ll need consulting service shops to achieve them, not development shops.

Consulting services partners only have linear contributions to your bottom line though. Managing and supporting them therefore needs to be a linear cost, and that implies keeping their toolkit minimalist and simple in nature.

The most elegant and efficient way to reach this state is to not provide a special toolkit to service partners at all; instead, partners work with the same toolkit that your own development teams use. Imagine a company in which every team’s functionality is available via service interfaces that designed to be eventually public. Such a company is not only using Conway’s Law for good, they are enabling partners by enabling themselves. This doesn’t eliminate partner-vendor squabbling, but it can keep the tenor focused on more easily resolved questions. It’s easier to answer “we want a bigger slice of these deals” (single variable, money) than “we want an easier and more flexible development toolkit” (what do easy and flexible even mean).

“APIs everywhere” as a partner service model also generates the maximum value for a development partner, who is now unconstrained. They may plugin to your stack anywhere and create value in any way. However, this is not an unalloyed good. Where many services partners are constrained to a single platform vendor (or at least a preferred vendor per use case), the development partner has a more flexible destiny. They are also more inclined to risk, since their business model rests on big upfront investments with uncertain but hopefully exponential rewards. If the platform vendor’s stack is completely open, a development vendor can easily subvert the vendor’s intention, and is far more likely to try it than a services partner. A few interesting examples: Elastic’s fight with AWS, AirBNB’s uneasy relationship with listing indexers, and Twitter’s on-again-off-again stance towards third parties. One might use an analogy: services partners for dependable, steady growth, development partners for high risk, potentially explosive growth. This can be a helpful model in deciding what vendors to support, but isn’t as helpful when deciding what toolkit to ship to them.

It’s worth picking apart the difference between technical support of a model and legal support of a model. Open APIs as a technical choice is a clearly beneficial system: internal and external teams are on the same footing, allowing maximal value to customers for minimal effort expenditure. The downsides of the model are in business risk. Remediation of that risk is a business problem, and the resolution is a partnership contract requirement and a technical enforcement via access keys. That’s obviously not an option for a fully open source system, but I can’t say I’d advise a fully open source approach to any platform business anyway.

Licensing models, self-service style

In my other two posts about licensing, I suggested that flat rate pricing is best for customers, but impossible in enterprise sales because of the variable and high costs of making a sale.

Those costs are difficult to understand if you haven’t been exposed before, but they are all too real. Weeks spent in negotiating a price are only the start; weeks spent in negotiating contract language are just a feature. What about indemnification? Can the vendor insure the customer against potential supply chain threats for the foreseeable future? It’s simply a matter of cost... and that insurance policy is now part of pricing.

What will happen to the deal if the vendor is purchased by another company? Can the customer audit the vendor’s source code? If the vendor goes insolvent, does the customer get to keep the source code? Yes, I have seen a customer organization running their own version of a formerly commercial product a decade after the vendor threw in the towel.

I was once involved in a contract between two industry titans that included a minimally disguised barter of services, and one of those services was sold to a third company as soon as the ink was dry. The cost to make and then keep that sale was... not small.

Even when it’s not titans you’re selling to... you can still be blocked on your ability to cross the competitive moat around enterprise software. If the thing you’re selling is close to the customer’s mission or has visibility across the customer’s entire org, they’re more likely to apply scrutiny and it’ll be harder to fudge compliance and legal details. The amount of blockage is directly tied to the amount of coverage or visibility your product will have for the customer. For instance, you might expect a gigantic financial customer to care greatly about indemnification, but they probably don’t for a specialist tool that gets quietly used by 25 people in the security ops center every now and then. Whereas if you’re selling something that sits in a mid-sized retailer’s cardholder data environment and manages the entire cashflow, they’ll probably care a lot more.

So as a vendor, there is a reasonable pressure to force your cost of sale down, and there is a clear goal: the almost zero cost clickwrap contract. Simply set your terms, disallow negotiation, and let the dollars roll in. It’s the ultimate expression of flat-rate pricing.

This is a fine approach for what I like to call lifestyle businesses: if you just need enough money for you and your cat to live happily, then sell away. The catch is that the most lucrative potential customers literally can’t buy from your business because of the potential risk. You’re probably good to go if your addressable market is consumers and your price fits on a credit card, but big business is off the table.

Wait! Singleton users and small teams buy in this model all the time! Expense report reimbursement is open to question, but no one cares if the price is low enough. A frustrated employee may just eat a few dollars for a productivity enhancing tool. The clickwrap model gets extremely blurry around personal computing appliances. I’m writing this in Bear on my iPhone, how is my employer to distinguish it from work I do with and for the company on the same device with the same app? (In my case, I use different editors for different roles.) Corporation and government legal departments try to draw a clear line, but IT struggles to implement that line and a clickwrap vendor is therefore always in danger of being pinched by changes in policy. Shadow IT is no place to make big money.

However, shadow IT does have some astounding success stories: Amazon Web Services is the obvious example, but Balsamiq, Basecamp, and Glitch (FKA Fog Creek) come to mind as well. If the official channels cannot support a use case and the need is great, then people will find a way.

Part four.

Sunday, August 25, 2019

Put PICA on Notable Events

For every notable event, the analyst adds a little PICA.

What’s a notable event? It’s a record that something happened, or an alert that something is expected to happen. It theoretically requires some form of response, from “read and move on” to “read and acknowledge” to “follow this run book” to “alert the [managers|Red Team|President] and [start the clock|increase logging|take cover]”. A notable event may be an Incident or Event in ITIL terms, a Ticket in bug tracker or fry cook terms, or simply grist data for a machine learning mill.

What is PICA? An acronym borrowed from the Dallas News by Clayton Christensen.
* Perspective: what is the importance of this event to the organization’s goals? Does it affect security posture? A service level objective? Is it a compliance breach?
* Insight: what is the cascade potential for the risk represented by this event? this event? Does it require immediate remediation or is it just a counter to be watched?
* Context: Is this event a one-off, or is it common? Is it more common for the grouping than the overall organization?
* Analysis: is this type of event occurring more or less frequently than in the past?

With a special incident, the statement is clearly true: The SAN is almost full. My perspective tells me that systems are going to stop working, and my insight into those systems lets me understand knock-on events across my organization. I know the context, why we need these systems to fulfill our mission and why that is important, and I use my analytical skills to determine a course of action.

However, every firewall rule triggered alert in a SOC or breakfast ticket in a diner does not immediately require a great deal of insight. As a developer, I see your low-impact typo ticket and I fix the bug.

There is still a need for PICA on these low-or-no impact notable events. Perspective: they still consume human attention, wasting the most expensive resource in the environment.  Insight: this kind of alert is ripe for automation, and a fine place to use a machine learning algorithm. Context: Reducing the flow of useless alerts makes important ones stand out better. Analysis: cost-benefit calculation suggests spending this much time to eliminate that noise.

Managing the Unmanageable

I’ve been thinking off and on about containers (FKA partitions, zones, jails, virtualized apps) and mobile ecosystems for a few years. These technologies have gone through several iterations, and different implementations have different goals, but there is an overlap in the currently extant and growing versions. Hold containers, IOS/Android, and MDM-plus-AppStore enabled laptops together and look at the middle of the diagram: 1: management is done in the surrounding systems, not in the daily use artifact. 2: management needs are minimized by simplicity.

A container is built, run, and deleted. There is no “manage”. To change or fix it, you go upstream in the process. A phone app may be installed or uninstalled, but it will take care of updating itself from someone else’s activities upstream in the process, just like a container. Users and admins don’t patch them, instead vendors push updated versions into an infrastructure that automatically does the needful. Even the infrastructure around the app or container, firewall policies, routing policies, device controls, all the policies and configuration that make the system secure and effective are also managed centrally and pushed into place.

This vision of abstracted management has attractions from many perspectives, which are obvious enough that I won’t waste time repeating them. It is also frustrating to teams tasked with monitoring and managing to existing standards of compliance. The new model is for computing appliances and services, and does not fit well with the current model of managing general purpose operating systems. It’s arguable if the computing appliance model can apply to general purpose computers at all; it’s theoretically possible to lock one down sufficiently but the result isn’t better than a mobile device. This attempt failed in the BYOD (Bring Your Own Device) laptop cycle, but the idea of being able to add and remove “appliance mode” on a general purpose device hasn’t died and only time will tell. BYOD seems to be working just great for phones, after all.

The power of systems management tools comes from the philosophy of the general purpose operating system. Programs run with each other in a shared environment which fosters their working together to serve one or many users. Users, including administrators, can remotely do whatever they need via networking. In the primordial slime of the business opportunity called systems management, administrators would use remote shells to script their desires into being, pulling packages into place when needed. Much has changed, but the fundamentals of these tools remain the same: a remote program with privileges, command and control networking, and a file movement tool.

The new model does not allow these fundamentals. We aren’t running as root in the remote host anymore. While mobile and laptop systems retain broader abilities, in the strictest container models even communication and files are only allowed to come from one place. There are exceptions as a matter of theory, but organizations who embrace the philosophy are going to prefer blocking those exceptions. And they will be right, because running visibility and control agent programs in a container or a mobile app sucks. Not only does it increase the weight and computational complexity of the target, it does so for no good reason; the fabric and philosophy of the new model are designed to prevent anything useful being done from this vantage point. Your process is not suppose to worry about other processes. As a user, you’re supposed to worry about your service fulfilling its purpose, not management functions.

This philosophy is not a comfort to compliance auditors, some infosec teams, or traditional systems administrators (hi, BOFH and PFY). It sounds too much like developers sitting in an ivory tower and announcing that they have handled everything just fine, a priori. Even if they say “devops” and “SRE” a lot. But at the end of the day, organizations are regularly accepting a similar statement from their everything as a service vendors, and not many can fully resist the new model’s siren song. But, a new computing model is not able to ignore law, finance, and customary process. The result is a grudging middle ground of management APIs, allowing a minimum viable level of visibility and control into the new model.

These APIs do not restore management fundamentals; they only allow you to log, to measure states, and to initiate change within the new model’s parameters. Posit that breaking the new model rules is going to fail, immediately or eventually. A management vendor is therefore in a jail cell, and has to differentiate from inside when offering visibility and control for computing appliances. Windows CE was the last gasp of general purpose operating systems for appliance friendly use cases (Linux may appear to be an exception, but the deployed instances used in appliances are hardly sporting full Unix shells). From here out, endpoints are full general purpose machines, a mobile approach, or a handful of frozen kiosk and VDI images. Servers are a mass of general purpose machines, mostly on virtualization, sometimes delivered as a service, with an explosively growing segment of service oriented app virtualization.

A new type of management agent is born for these API-driven appliance models. Maybe it’s implemented in “sidecar” containers or as “MDM approved” apps, or maybe it lives fully in the cloud, maybe it’s the focus of a new vendor or the side project of an established one. There will certainly be pronouncements that it brings new value to the use case. Doesn’t matter how it’s implemented or marketed though, it’s accessing the same APIs as everyone else. Its best efforts are limited to “me-too”. Differentiation is either in costly and difficult up-stack integration, or a capital-burning race to open sourced commoditization.

A customer who wants single pane of glass visibility, is left with few options: build their own analytics, invest in data lake technologies, or buy extensions to their main management tools. Almost all select two of the three for resilience.

It may make an unpleasant experience for the management tool, where this ghost of management is fit into the same console and mental model as a full-powered vendor’s real capabilities. “Here is your domain, in which you can do what is needed to ensure your organization’s mission! Except on these special systems where you know a lot less and can’t do much of anything.” Customer expectations are sort of hit but kind of missed, and no one is very happy. Some vendors can sell “know less and do less” alongside “full visibility and control” for the same price. Others may adjust the license model instead.

So, is the single pane of glass worth a cognitively dissonant user experience? Or does the customer split their visibility and control tools and buy something else to glue things back together, moving that dissonance higher up the stack? Because there will surely be dissonance when clicking for action in tool A has to go through tool B’s brokerage into Tool C for execution.

There is a useful comparison to minority or legacy operating systems. Management and visibility tools universally reduce their capabilities on platforms that aren’t as important to their customers, so very few are excellent on Solaris, AIX, or HP/UX. The important difference is that a vendor’s reduced AIX capabilities are a matter of choice. If the market demanded, the vendor could eventually resolve the problem. A management vendor cannot change the operating model of an entire ecosystem, so computing appliances are not like legacy computing. But there is an analogy in that the tools do not align perfectly with customer needs, leaving gaps to fill with people and process.

If we imagine a perfectly amazing management tool for AIX that doesn’t integrate with the tools used for Linux and Windows, the choice becomes clearer. Customers don’t require visibility and control for operating systems or computing models, but rather for business functions and services. Buying different tools for different systems can be a required stop gap, but it’s not a goal in itself. Therefore, a single product, single pane of glass approach wins over a multi-product, best of breed approach. The remaining question is therefore one of approach: do you use an endpoint-centric vendor that was born from visibility and control, or a data-centric vendor that was born from searching and correlation? The answer lies in your organization’s willingness to supplement tools with labor. A data lake can have great visibility, but it has no native control, meaning another gap to cross before even hitting the API gaps in the new computing model.

The goal of the new model is to minimize and ultimately remove management entirely. As long as it is unsuccessful in this goal, there will be rough edges between the new model and the old. Those edges bias towards the old model consuming the new.

Sunday, August 18, 2019

Everything I know about the IT business is from Godzilla Against Mechagodzilla

Which movie is this? There’s a few, so it’s important to disambiguate!

Not to be confused with (which I can only recommend to the fevered or otherwise hallucinating) or even (pretty fun, but not a sufficient scaffold for understanding one’s career). To be fair, would probably also work for this exercise. Let’s get started!

When problems become apparent enough to need resolution, they’re often all-consuming.

Your old plans for resolution are for your old problems, and might not work any more.

Someone or something is going to take a fall. New attempts to fix this culture are laudable and I hope they succeed.

A new solution is much more likely than fixes to the old solution.

People struggle to put rationality ahead of emotions. This fellow blames people instead of technology for past failures.

New solution projects always have surprises and grow to cost more than anyone expects.

They also still go into the field untested against outside context problems...

... and bad things can happen when testing in production.

Metrics data is fine, but user interfaces should always put the bottom line up front.

Give an analyst an alert, and they’ll want all the data.

Someone always ends up taking one for the team.

Stated differently, if everyone only performs to the stated requirements then the project won’t be successful.

Inelegant solutions need a lot of power.

Partial success is better than complete failure.

Analysts will grow fond of their tools, even if the ultimate outcome was only partially successful.

Tweetise here:

Friday, August 2, 2019

Enterprise Business Metrics

So you’ve launched a product... Is the product selling? How’s ASP (Average Sales Price) after discounting? Deal size? Cost of sales? Are there measurable predictors for losses? Are there ways to accelerate or increase the wins? If your license is per term instead of perpetual, is your product getting renewed?

And if your product is one of many...  are your wins correlated with the wins for other products? What would happen if they were bundled? Does your product cannibalize something else the company sells? How would you know? Do ELAs hurt or help your product’s adoption?

Companies ask these questions because they need to manage the business. Tautology, right? So let’s be blunt: you can’t get investment or spend investment without some way to predict how you’re doing, and you can’t decide if you’re going to continue an investment if you can’t see how that investment is performing. As a product leader, your paycheck is coming more or less directly from that investment and these questions have a ring of immediacy to them.

And if your product is a consumer-facing direct sales widget, you may be facing some mysteries (aren’t we all), but the numbers are probably relatively clear. Unless it’s sold through retail partners. Enterprise software sales though... when your product starts at “new car” and can cost up to “Central Park penthouse”, measuring performance starts getting strangely difficult.

Enterprise software is sold to customers who don’t always want to be clear about how much they are willing to pay or when they are willing to pull the trigger. At the very least, this means that the deal data is unclear and may change for reasons that don’t involve your product.

Enterprise software is sold by sales people, and sales people are maximizing their compensation plan and pipeline. At the very least, this means that entering the data you want into Salesforce is pretty low on their priority list. At the extreme, it can mean a variety of bad behaviors, particularly if the sales person is not actually competent. They might have good reasons for sandbagging deals or inflating pipeline, or they might have heard it was a good idea somewhere and misunderstood the reasons... but all sorts of craziness can happen in an enterprise sales team.

Setting aside the truly bizarre behavior of a failing team, sales leadership might try a bunch of mechanisms to deal with the normal lack of clarity. Favorites include:
* Dedicated people who force the deal to make sense. They might be called something like “sales operations” or “deal desk” or “contract specialists”, or the function might be overloaded onto an inside sales team. The resulting organization is simply fatter than before, because there’s still customers and salespeople with their own motivations and context in between the data and the organization.
* Punitive policies: the deal won’t be booked or the sales person won’t be paid if all the reporting isn’t done in a correct and timely fashion. This is an amusing game of chicken because the company willing to a) not sell product or b) risk a lawsuit over a principle of report quality has got their priorities seriously backward. Actually going through with such a threat is a great way to lose customers and sales people, which really reduces your sales numbers.
* Rewarding policies: the sales person will get a toy or points toward the yearly club or public recognition for doing their reporting in a correct and timely fashion. Again, simply amusing, because this data is not worth an incentive large enough to motivate a sales person worth hiring. A good sales person in enterprise software makes very large amounts of money. You can play on their sense of camaraderie and you can ask them to be diligent, but those factors are present without the incentive. In order to incent, the prize has to mean something in relation to their compensation, and that number is not small. Furthermore, the sales person’s job is to make the customer’s organization complete the purchase, and the incentive has no impact at all on the customer. Is it supposed to make the sales person work harder? Then why isn’t it simply paid to them as part of their total compensation? Are they supposed to give the incentive to the customer’s purchasing department? Sorry, that’s illegal bribery.

Given the ineffectiveness of these interventions, why do companies pursue them? Any generalization will miss a lot of examples, but I am fond of two explanations: the manager who is more comfortable with spreadsheets than conversations, and the manager who isn’t sure what to do, so they do something that they understand.

So we return to thinking about what the deal data is worth... there’s a quote popularly attributed to Charles Babbage, “Errors using inadequate data are much less than those using no data at all.” If the data in Salesforce can be considered directionally correct but untrustworthy in detail, it is still useful for the purposes listed above. You can manage with it. You’re working with a string and a rock instead of a titanium yardstick and a laser level, but a lot of buildings have been erected like this. Improving the quality of your sales measurement tools is worth very little when compared with double-checking their results: talking about specific deals with sales people and customers can be remarkably illuminating.

Working with a Coach

Executive coaching and mentorship is an interesting part of modern business, and sometimes people are not prepared for taking advantage of it. Here are some notes on the purpose and value.

So you’re working with a coach to get better at execution... what are you going to say?

As a potential mentee, start with outside boundaries, the Overton window. A corporate coach is not a therapist; the need is to find performance problems and optimize teamwork. More Dave Righetti, less Sigmund Freud. You may uncover reasons to work with a therapist, because leadership work involves you as a person far more than individual contributor work does. Your emotional resources are going to be used when you lead, inspire, console, and support others. The coach is not the person to help with personal maintenance, but they might be able to show you what you need. Fixing those problems is work to do elsewhere.

Within what you’re willing to discuss, locate the secrets. Why don’t you want that opinion known? Is the reason rational and explainable to an outside coach? Is that reason a you problem or an organization problem?
Ideally you’re not trying to discuss anything that your team couldn’t already know, and the question is how to be more effective in communicating your thinking, persuading others to agree, or accepting that your opinion didn’t prevail and the team is going elsewhere. If the situation doesn’t look like that, there’s something to fix. Is it you, or is it the organization? Is it reconcilable, or will it end in parting ways? Might as well focus on figuring that out before you bother with unpacking whatever secret has led to this realization.

Now, back to the coach. This person fits into one of three boxes: A mentor who is helping you for charity and their own growth, a coach that your organization paid for, or a coach that you paid for.

Mentors may or may not be encouraged by the organization, and there may or may not be a formal mentorship program; in my opinion those parameters only affect the degree to which the mentor is actually willing and able to help. A mentor who is able to see problems and help you fix them is helpful, regardless of what structure the organization provides or does not provide. Notably, a mentor’s time is limited and you need to aim for efficiency. If you don’t have a specific and achievable improvement goal in mind, you’re not getting mentorship, you’re having coffee with a coworker.

The existence of an organization-paid coach may open some questions: why is this person assigned to me? Are they being hired for others as well, or am I singled out? You may feel like you’re being prepared, for greatness or for unemployment. It’s best to assume this is a positive investment on the organization’s part in order to make you (and potentially others) a better leader. There is little to be gained from indulging in paranoia. However, it is fair to note that a person hired by the organization has loyalty to the organization as well as to you. You should think, and communicate with the coach, about what information you are willing to share. Again, time is limited, but you probably don’t know what the budget is; it’s best to get what you need fast.

A coach that you pay opens a different kind of risk: they are potentially seeing the organization through your eyes alone. Even if they’re able to attend your meetings as your guest, they are always getting a filtered experience. Ideally the fact that you’re paying will help you focus, but that’s not always the case; you should plan for a fixed number of sessions up front and clearly state your immediate and achievable goal.

A good coach will see small and large blockers, and they’ll sound trivially obvious. You may even already know you have these issues. “Discussing the problem and constraints first is good for engineers, but bad for executives; start with your proposal instead of how you got there.” “Don’t cross your arms and look away when you’re asked a question, you look defensive.” It’s not helpful to discuss the existence of the blocker; move on to their recommendation, and sincerely try that recommendation at least three times so you’ll know if it was the right thing.

Bad outcomes can occur in mentorship and coaching, but by far the more common result is an ineffectual series of meetings. If the mentor is unable to provide helpful advice, or if the mentee is unable to act on the provided advice, or if the actual problem is outside of the mentee’s control, then coaching is really only relevant to a future situation where the mentee is hopefully able to use what they learn.