The new paradigm of productivity

Image Credits

A brief history of modern productivity tools (PT)

We can view the evolution of productivity tools (PT) in 5 waves — each leading to a substantial improvement over the previous.

• Wave1: Physical → Digital: Modern PTs started in late 1990s with Microsoft offering a suite of products that digitised existing systems (like physical diaries, to-do lists). This combined with the WWW unlocked several productivity benefits (e.g. immediate visibility and ease of communication around meetings [accept, reject, propose time], reminders, etc.). At a high-level, it led to “time-savings” and “convenience”
• Wave2: Cloud-hosted/Multi-player: Google spearheaded the next wave in 2006 with it’s launch of Google Calendar & Google Docs which were cloud-hosted, making them easier to collaborate on. These multi-player tools helped significantly reduce the back-and-forth on scheduling (by making calendars shareable) and versioning effort of working on and collaborating on offline docs/sheets/slides (by making them collaborative). This led to a substantial “time-savings” and “convenience”. Note: Figma, Notion etc. also belong to this wave.
• Wave3: Joyful Productivity: The last 4–5 years has seen the rise of the “joyful productivity” movement led by the likes of Superhuman & Amie which focuses on helping users attain a “state of flow” while working. They achieve this through higher load-speed, ultrafast navigation/action through intuitive keyboard shortcuts, powerful triage and a better UI (esp. gamification).
• Wave4: Smart productivity: While Wave3 focused on building a better workflow to enable a state of flow, the latest paradigm tries to achieve it by reducing the mental load by automating things that break that state of flow. e.g. Mem reduces the mental load of organizing notes, tasks & other sources of knowledge by leveraging AI to build a powerful knowledge-graph of all the info stored; e.g. Clockwise reduces the mental load of determining when to schedule a meeting by automatically adjusting internal meetings to optimize free blocks of time for deepwork.

Note: It is difficult to establish the chronology of Wave3&4 because they seem to be somewhat overlapping in timeline & also features (e.g. “smart triage” belongs to both 3&4). The result in both cases is “time-savings” and joy/flow (Superhuman is often quoted as a “wellness product” by many)

• Wave5: Smart + Creative Productivity? As we look into what the next paradigm of productivity could be, AI PT use-cases advancing from “organizing/optimizing/automating workflows” to “Creating”. e.g. Flowrite improves your ROI on time by creating an email based on just inputs around “intent”. Important to note that while “time savings” is still an important byproduct, the new wave might also bring along advances in “quality improvement”.

👉 As these tools improve, the expectation from knowledge workers grew at pace or even faster.

Key Areas of focus

As we deep-dive into productivity tools, it is important to structure the key areas of focus — largely determined by where knowledge workers spend the most time.

• Communication: Email, Slack, IM (Superhuman, Flowrite)
• Time Allocation: Calendar + Tasks (Calendly, Clockwise, Amie)
• Knowledge Consumption & Storage: Browsing + Note-taking (Mem, SigmaOS)
• Organization & Retrieval: Folders, meta-tags (Dala, StayTidy)
• Collaboration: <let’s keep it out of scope for now>

SignalFire’s Future of Work taxonomy is a useful starting point to further break-down the productivity tools use-cases.

Drivers of change in this space

There are a few long-term trends, which will influence the next wave of productivity. Most of these drivers seem to be quite permanent from a 5–10yr perspective and worth focusing on.

ML/AI have come of age

Advances in ML/AI are expected to be a leading driver of change in the PT space. While they currently optimize workflows (triaging, observability, contextual search); Generative AI now have the ability to create (with GPT3, GAN). This can unlock a new wave of productivity — one where users need to focus just on decision/intent while the tool takes care of execution while understanding the context — this is hyperleveraging time to a whole new degree. Another potential impact of this will also be on “quality” of the output (e.g. salespeople will be able to create 10x faster&better cold-email/replies, IT companies would be able to respond to RFPs/Tenders 10x faster&better)

Hybrid work is creating a deeper digital footprint of work

As people increasingly turn “hybrid”, their digital footprint is expanding rapidly, as their reliance on SaaS tools grows. An even more interesting phenomenon is that the digital footprint is also “deepening”, and our SaaS tools have much better context now of the shared organizational-brain than ever before (Lot of content lost in physical meetings is now happening digitally [meetings → loom, zoom][water-cooler chats → slack channels). This has two implications:

1. A need to organize this big-data and make it available seamlessly. Further, a lot of this big-data is not text, but rather in multimedia format (images/video/audio) which NLP/STT/Image-Recognition are now able to understand much better and at scale
2. An opportunity to leverage this big-data to optimize/improve our way of working

Geographically-remote work demands fundamentally new ways of collaborating

Geo-remote work, with multiple timezones and lack of physical colocation is challenging the existing processes around collaboration: daily-standups/huddles, whiteboard sessions. While v1 of the solutions trying to solve this have focused on enabling these processes to be conducted online, there is a real need for fundamentally reimagining the workflow itself.

Further, since the patterns of working will vary so widely across all employees, it would be difficult to find solutions tailormade to workflows of small remote teams. In addition, it would be inconvenient and expensive to work with central engineering teams to build these workflows for every single team. This might drive organizations towards self-reliance and personalization of their personal & team’s tech-stack with no-code automation.

Purchase decided/influenced by the end-user

A key trend, expected to continue over the next half-decade is that the purchase decision for productivity tools has shifted from the CTO/CIO → executive → end-user.

• This has led to the rise of the “power-user”. Instead of tools being built to satisfy the “least-common-denominator” of users, they can now be built for users who might be able to see 10x higher value in using them. Such users not only are willing to pay a premium for them, but are also willing to invest the time to get onboarded onto the platform (incl. learning the workflows, keyboard shortcuts, …). One should expect more ‘point-solutions’ like superhuman become viable due to the power-user phenomenon.
• This trend is also leading to a rise in standards of the individual product being used — with increased flexibility on their tech-stack, users are less willing to compromise on features/workflows/interfaces and more willing to try out new products to meet their needs. This implies that new tools need to have high standards on not just functionality but also UI/Design on day0.

Key Opportunities in this space

ML/AI will move-on from “just time” to “quality” & “access” (Flowrite)

There is an opportunity for new “create” tools that in-addition to saving time, also help improve the quality (e.g. of sales emails) or provide access to services which were unafforable earlier due to it being manually intensive. In addition, these new tools can also help not just optimize, but also “personalize” work — e.g. Flowrite could learn how to write emails in my style, by learning through my past emails.

The rise of the robot personal assistants (Saiga)

Another example of improved access is “robot personal assistants”. As EAs are inaccessible to the everyday knowledge worker, we solved for it through self-serve tools like calendly (scheduling), opentable (for booking restaurants), etc. Now, it could be possible to have a new type of robot personal assistant who can be accessible at a much lower cost (Saiga starts at $299/mo). Instead of being powered by a better TTS/STT & NLP tech (Alexa), this will be powered by a strong decision & action engine at the backend.

Knowledge Management will not be through “folders” (Dala, StayTidy, Glean)

The old paradigm of folders is being challenged by the multitude of platforms (how do you replicate and make the team follow the same folder structure across notion, gdrive, airtable, …) and multimedia (how do you consistently as a team tag all images, videos, audio — including communications on slack?) being used for communication now. The answer to this is AI/ML driven organization of content & contextual search across docs/files/videos/comms/browsed-websites/call-transcripts.

There is value & potential for a [J.A.R.V.I.S](https://ironman.fandom.com/wiki/J.A.R.V.I.S.) — like product that can answer questions like “what was that blog-post on PLG that I read sometime last week?” “Can you pull-up the loom where we discussed the product roadmap?” “Who should I talk to in Moonfire regarding the modern data stack?”.

Minutes of the Meetings no-more (Eric.ai)

There are already several “meeting transcribing” apps for zoom and g-meet; but there is a potential to take them to the next level. New tools could not just transcribe the meeting, but also take notes, understand the key implications and write the minutes of the meeting (something that junior team members painstakingly spend upto 30mins doing for every meeting). Not just that, it could take that to the next level by linking directly to Jira and creating a ticket for every task in the MoM.

Productivity tools will be built for the Niche Power-user Personas (Superhuman, SigmaOS)

• Users will be willing to pay for a 10x better performance
• These might increasingly become “vertical/functional” solutions

Intelligent No-code/Low-code (NCLC) automation/ Layer2 RPA (mimica, Airtable/Stacker+Zapier)

Another exciting opportunity stemming from (a)deep-digital footprints and (b)heterogeneity in team workflows is No-code/Low-code automation of team workflows. As teams will struggle to find the right off-the-shelf tools to automate recurring tasks in their workflow, they will increasingly turn to RPA. While org-wide critical & high ROI workflows would already be undergoing RPA, smaller teams might be left to figure this out on their own. That’s where non-technical business teams might leverage the power of NCLC automation. Use-cases could traverse the familiar trajectory of reporting → observability → non-critical actions → critical actions. Another potential area on opportunity might be “intelligent NCLC” — where the deep-digital footprint of the organization is analyzed by AI/ML to identify potential sources & pathways to automation. We could call this Layer2 RPA.

The introduction of “Wrappers” to integrate the 100s of SaaS tools/employee (Qatalog, Omnifia)

Proliferation of point-solutions has also created enormous fragmentation for organizations and individual employees. In one day, an individual might have to track comments/pings from 5+ tools (slack, figma, miro, whatsapp, email, gdocs, notion, …) multiple times. There is also no way to quickly onboard someone on “what is done where?” in the team. This might pave the way for “Wrapper” solutions to integrate and establish “ways-of-working” with these SaaS tools.

The Winning Hand: what should early investors look for in opportunities

1. Founder with PLG experience: PLG requires a concerted effort from product/biz/marketing and the founder is best positioned to bring this together. A founder with prior PLG experience would be a great enabler for a successful PT startup. While this is especially true for point/end-user solutions, even for team/enterprise products, the primary mode of sales is likely going to be bottom-up and PLG (e.g. Qatalog) — making PLG experience almost a blanket requirement. A corollary to this, is having a founding team with a strong design mindset. As these tools become easy to try and easier to discard (point solutions, easy data migration, connected systems), even MVP1 needs to have a strong design. Studying some of the more successful tools in this space also hints towards this (Ivan of Notion spent 18hrs+ a day on figma, Rahul Vohra of Superhuman was vocally obsessed with the UI)
2. Community is super-critical: A key theme across all top productivity tools, has been the community around it — notion had a fledgling subreddit since inception (200k+ users and growing very fast), mem.ai had users generating videos on how to optimize their notes/tasks using mem even as they were in limited Beta. While the medium depends on the persona of the early niche they are going after (slack, reddit, discord, twitter), having a community element is a critical part of their strategy. Even at pre-seed stage, when it would be premature to expect a community already in place, founders having strong views on how they will engage community will be very important.
3. Monetization Proof: While a lot of productivity apps manage to get an initial critical mass of users (ProductHunt is littered with those examples), they usually struggle to convert their early adopters to paid users. Often the users don’t value the “premium features” as much, or don’t value even the core product enough to pay for them (Several calendar apps faced this challenge — Cron sold to Notion in May ’22, Hera folded in June ‘22). Hence, being able to see a roadmap to monetization is critical for a winning company. Early-proof of actual paid users is necessary if it’s a power-user 10x experience play. Power-users should be willing to pay from day1, which is the most conclusive proof that it might be a 10x better value-prop (counterpoint: calendly started as free). Another strategy adopted by the likes of Notion, mem, is to let individual users continue to use it for free, but charge for “collaboration” use-cases. Most pre-seed/seed companies might not be at a stage where the collaboration workflows are fully built out — in such cases, having a clear logical roadmap & rationale to a deep collaboration-play will be vital. (mem.ai has evangelical users for their individual knowledge graph, but is struggling to convert it into a team-knowledge-graph)
4. AI/ML tools need to have their own powerful proprietary data & engine, and not just good UX on a commodity engine:
   e.g. Saiga’s proofpoint is ability to perform tasks; rather than whether it can take NLP commands
   e.g. tonnes of auto-scribe apps for zoom which all use the same NLP & STT kits, lack defensibility and are unlikely to grow without meeting a lot of competition. Individual usage-driven proprietary data is probably the most scalable and defensible data source, and hence we need to have an open mind about companies that use initial distribution hacks to gain the early-critical mass necessary to access sufficient data.
5. Combination of [Tough learning curve for customer + seamless onboarding workflows]: If a product is sufficiently complex that it has a steep learning curve for the customer, it can offer a big advantage and defensibility against new entrants. Winning startups achieve this in the initial stages by: (a) Have a superb onboarding process — even if slightly unscalable at first (1-on-1 onboarding at Superhuman, “always on” chat for Notion) (b) When sufficient early users have started using the product deeply, transition to a self-serve onboarding flow which. Focus on simple use-cases which are easy to learn and communicate initially (Notion uses GIFs & “replaces A,B,C”). Prioritization based on “Immediate Time-to-Value”. Gradually introduce more complex use-cases (e.g. daily onboarding emails by Superhuman/ Notion in first 2 weeks). Lean on power-users/community champions to support new users. If product is in initial stages, don’t worry too much about #Users or Brand of Design Partners but rather about depth of engagement (e.g. if good engagement in small set of users, at least it’s a 10x product).
6. Scope: logically lowest building block: Productivity Tools could be viewed as building blocks in the overall individual/team tech-stack, and hence that implies the best product for a narrow-use-case wins (easy to replace, easy integrations, individual purchase decision). Hence, the aim should be to win in the logically lowest building block. Important to note that this might not necessarily be the lowest possible block. e.g. Calendars and Tasks are the lowest possible blocks/primitives for a tech-stack, but a logically lowest building block might be Calendar+Task with collaboration features so that they tightly talk to each other (e.g. Amie). Hence, the scope needs to be large enough to be logical (e.g. calendars + tasks, email + some scheduling element, notes + tasks + databases, …), but small enough for a startup to be 10x better at.

Interesting Reads

1. David Ulevitch on “The Developers’ Way”
2. mem.ai founder on the new paradigm of knowledge management
3. First1000 on how matter, notion, superhuman, calendly etc. got their first 1000 users
4. Immediate Time to Value — another way to scrutinize PLG companies
5. a16z Superhuman announcement
6. a16z mem.ai announcement