#68 - You Don't Have a Strategy If You Don't Say No
Plus: Hybrid work & Distributed teams fail slowly; Success profiles for jobs
I think I helped a team find the courage (and the organizational support) to say “no” to things this week.
Strategy for teams is hard. I’ve sort of given up on trying to find strategy articles for the roundup that are suitable for us; most are for either large organizations with many moving parts which make no sense in our context, or for tech product strategy. The product strategy ones start off okay, but they really imagine being able to pivot products to very different markets, and that doesn’t work well for our teams in larger organizations.
The best articles I can find are for consultancies or for nonprofits or even new PIs, where the advice is the same: find a niche at the intersection of (things your team can excel at) and (things that are needed and not provided well elsewhere), focus relentlessly on that, and shift or expand focus only reluctantly and judiciously. Crucially, the niche should almost certainly not be an internally focussed, technical one but an externally focussed, client-centric one. We do want to find product-market fit, it turns out, but the product is our team.
With a little care you can handle a portfolio of such niches — it muddies communications a bit, but it’s certainly something that I see teams manage. But having none, or too many, makes success unreasonably hard. The one thing that all strategy discussions have in common is that a successful strategy is defined in part by what you won’t do, by what you choose to be out of scope. Absent clarity on that, no one knows exactly what you do (in any language they care about), and the typical failure mode in that case is trying to be all things to everyone. In that case you end up doing poorly at some of those things, soiling your reputation, and starting a vicious downwards cycle.
In this week’s case, an external team asking me for advice had been struggling for some time with a lack of clarity of mission. It was clear to literally every observer in what area the team was most successful, and where there were modest growth prospects, but culturally and organizationally the team felt it needed to accommodate every single request that came in, especially when money unexpectedly became tight. The result was flailing, high staff turnover, a wildly uneven reputation, and a larger organization that didn’t know what to do with it.
What they needed wasn’t a flash of strategic brilliance, or marketing genius. Literally no one was surprised by the outcome. It just took the pedestrian, labour-intensive work of making explicit that internal and external consensus around what the team should focus on, and building a consensus around what it was ok for them not to do. That meant re-casting what was the right scope several times, finding the right language to pithily summarize the scope, and pointing out the struggles (and duplications!) of trying to do work outside the scope. None of this was rocket surgery, it just took time and conversations.
Supporting research and your directs by being a professional steward of a research computing team isn’t easy, but it is simple. It is being deliberate about what you and your team are doing, and learning from what you and others have done.
Anyway, on to the roundup!
Hybrid Work: How to Get Ready for the Future of the Office - Mara Calvello, Fellow
We’ve all learned how hard it is to manage distributed teams in the past year. But most of us (hopefully) are going to moving to some kind of hybrid distributed/on-site work configuration when things come back, and by all accounts that hybrid mode is even harder. It’s really challenging to not have the distributed team members feel out-of-the-loop compared to on-site staff, and for on-site staff not to cut corners (documenting decisions and designs, etc) that we’ve learned are needed for distributed work.
The main thing now is to figure out (with team member input) what things will probably look like in the first instance, and start deciding what will be needed to mje that work. Calvello goes through three models:
(FWIW, our team will land somewhere between remote-first and office-occasional). In either case, she then talks about some of the work that needs to be done - the bits most relevant to our teams are:
Invest in team culture
Set clear and consistent expectations
Make sure remote employees feel included
Make hybrid meetings work - this seems like the hard part to me.
Research: Dispersed Teams Succeed Fast, Fail Slow - Marie Louise Mors and David M. Waguespack, HBR
Fast success and slow failure: The process speed of dispersed research teams - Marie Louise Morsa and David M. Waguespack, Research Policy 2021
As many of us consider a continuation of distributed teams post pandemic, some research suggests an issue I hadn’t seen before with distributed teams - they may be slower to admit that an effort failed and to move on.
In a study of 5250 already-formed research teams in particular, Mors and Waguespack found that given that the team existed and was successful (not a gimme!) the dispersed teams were already pretty good at working together and required less coordination and iteration to arrive at success:
In general, dispersed teams spent less time and went through fewer iterations before reaching success than the co-located teams. This suggests that team members are aware that it’s difficult to coordinate when they are located in different organizations ..
On the other hand, throwing in the towel took much longer:
Additionally, it may also be that it’s when failure occurs that the coordination costs kick in: Research also suggests that dispersed team members may find it more difficult to communicate around a failure or reach agreement on when it is time to abandon the project.
In the HBR article, the authors suggest a couple of interventions to prevent teams getting stuck on half-failed projects for a while - either more rigorous up-front screening to weed out projects less likely to succeed, or sync-ups or management intervention to flag that success doesn’t look likely and move on.
Increase your hiring success with job success profiles - Rod Begbie, LeadDev
TinyBird Tech Test - Javi Santana, TinyBird
This sounds like how hiring works at most of the research computing teams I’ve seen:
I’d seen enough job postings in my life to know you just had to come up with some qualifications (‘BS or equivalent in Computer Science’, ‘5+ years experience with Python’, maybe even a cheeky ‘Ability to work both independently and as part of a team’), throw them into a bulleted list, and start picking the best candidates from the resumes that will surely flood in
It’s hard to know if someone will be a good candidate for a job if you don’t know what success will look like for that role. Most people involved in the candidate selection and hiring process probably have a half-formed implicit idea of what that would look like, but they’re probably different. Unless it’s written down, it won’t be well though-out, agreed upon, and applied consistently.
Begbie suggests having clear primary objectives for the job, secondary objectives, and success indicators for 30 days/90 days/6 months/12 months. Not only does this clarify things internally and greatly increase the odds that you’ll be looking for the right people, by putting this information in the job ad you’re much more likely to attract candidates who actually want that job.
Relatedly, once you have a success profile it’s much easier to understand how to evaluate against that profile when you’re interviewing. Santana provides one real take-home problem they use at TinyBird, a company that builds real-time data processing tools. It involves writing up how you would solve a data ingest-plus-expose-an-API problem, and describes the rubric they use to answer it (it’s almost all about the communications, not the technical beauty of the proposed solution).
Navigating your filesystem too boring? rpg-cli turns every cd command into a move in a rogue like game. cd ~ — if you dare.