Continuous Improvement | Startup GTM Bible

By now we have metrics coming in and continuous improvement processes extracting insights. Phase 6 ultimately is about feeding those insights back into your GTM strategy and evolving it. A go-to-market plan is not a static blueprint; it’s a living strategy that must adapt to market feedback, just as a product evolves via user feedback. Dave Kellogg’s writing often points out that as a company scales, what started as founder intuition (who to sell to, how to pitch) must give way to data-driven refinement. In this chapter, we’ll explore how to iterate on three fundamental pillars of your GTM strategy in response to what you’ve learned: your Ideal Customer Profile (ICP), your messaging, and your channel mix. We’ll also cover how to ensure these learnings influence the product and high-level strategy, completing the feedback loop from market back to company direction.

Refining the Ideal Customer Profile (ICP) – Sharpening Your Target

Your Ideal Customer Profile (ICP) defines the types of companies (or buyers) that are the best fit for your product – those who find the most value, close relatively easily, and tend to stick around and grow. In early startup days, the ICP is often based on hypotheses (“We think tech startups with 50-200 employees need our solution the most”). By Phase 6, you have actual customers and deals to analyze, which means you can iterate and tighten your ICP based on evidence.
As Dave Kellog puts it: In the seed-stage, ICP is based on the founders’ vision; by growth-stage (say, >$10M ARR), “it’s no longer about vision, it’s about math”. You should have enough data to ask, who are our most successful customers and what do they have in common?.
Note, “successful” can be defined in various ways – part of the ICP refinement process is choosing the right success criteria. For instance, is your ideal customer one that closes quickly? (Sales team would love that, but if they churn after a year, not ideal.) Or one that deploys successfully and renews? (Customer success cares about that.) Or one that grows into a much larger account over time? Ideally, your best customers do all three: buy relatively easily, are happy and renew, and expand their business with you. Those are the golden cohort.
So how do you iterate your ICP concretely?

Analyze your customer data: List out your customers and key attributes: industry, size (employees or revenue), use case, source (inbound vs outbound), product features used, etc., along with outcomes like ARR, days to close, and retention rate. Look for patterns. Perhaps you find that mid-market financial services firms have the highest win rates and a 120% NRR, whereas small e-commerce retailers have low win rates and 50% yearly churn. That’s a flashing sign of where your ideal fit lies. It might not be so stark, but even subtle trends help. You can do a regression or scoring model that correlates attributes with success criteria. For example, you might discover companies using a certain complementary technology (say, Salesforce CRM) have much higher success with your product (maybe because your integration with Salesforce is a big selling point). That attribute can be added to the ICP definition (e.g., “Ideals: must be Salesforce shop”).
Re-evaluate success criteria: As mentioned, define what “ideal” means. Kellogg lists questions: big initial deal vs. low cost to acquire vs. high NPS vs. long-term retention – which matters most? Ultimately, the company wants customers who renew and expand, not just ones that sign big once. So you might weigh retention and expansion more heavily. If one customer paid a ton but was a nightmare and churned, don’t let the large ACV fool you – they weren’t “ideal.” You might rank customers by a composite score (e.g., Score = ARR * gross margin * expected years lifetime – acquisition cost, or something akin to LTV minus CAC) to truly find the highest value customers. Those are your North Stars for ICP.
Re-segment if needed: Often, initial ICPs are broad (e.g., “We target SMEs across any sector that need HR software”). Feedback might reveal that certain segments within that are much better. For instance, you learn that healthcare companies have special compliance needs your product handles well – they love you – while retail companies churn because they needed more customization. The smart move: tighten the ICP to “healthcare SMEs with 50-500 employees” perhaps, and de-prioritize retail. This can feel counter-intuitive (“we’re narrowing our market!”), but focusing on your sweet spot can dramatically improve efficiency and win rates. You can always expand to other segments later when product/fit improves there. The continuous review of GTM performance might show segment-wise differences – heed them.
Identify patterns in the sales cycle: If part of being “ideal” is also a smoother sales process, examine that. You might find your sales cycle is 2x faster when selling to companies below 200 employees than to those above 1000 employees (no surprise, enterprise deals take longer). If you’re a startup needing faster cycle, you might bias your ICP toward the smaller end of your range for now. Or conversely, maybe enterprises take longer but convert at a higher rate and have way larger ACV – maybe that’s worth it if you have the runway. Align ICP with your strategic goals (fast ARR growth vs. large strategic wins vs. land-and-expand potential). The key is you’re deciding based on data, not guesswork.
Iterate with cross-functional input: Bring in Marketing, Sales, Customer Success to discuss ICP tweaks. Sales might say “We keep closing deals with manufacturing companies, but they always churn because they need on-prem deployment which we don’t do well.” That’s vital info – maybe manufacturing shouldn’t be ICP until product changes. Or Customer Success says “Our happiest customers are those who use Feature X extensively” – check which types of customers use Feature X most, and refine ICP around that use case. Ideal customer definitions can evolve from broad (“industry/size”) to more nuanced (“companies that have problem Y and value proposition Z resonates, often characterized by...”).

“The ultimate ICP is built around customers who (1) we can sell to fairly easily, (2) who don’t churn (they renew), and (3) who expand over time. That’s what marketing needs to go get!” - Dave Kellog
That’s a great north star to steer by. So if your analysis and experience in Phase 6 point you toward a subset of customers that meet those criteria, consider officially updating your ICP documentation to focus there. For example, you might narrow your target from “all SMB finance teams” to “mid-market fintech companies who need SOC compliance (because they close fast and stick around).” This then cascades into marketing targeting, sales prospecting lists, and even product roadmap priorities (build more for those ideal users’ needs).

A Hypothetical Case Study

Let’s illustrate with a hypothetical case study: DataSync started selling a data integration tool broadly to any industry. After a year, they find their best customers by far are in the e-commerce sector, using Shopify, who need to sync order data to their ERP – these accounts onboard quickly, have low support needs, and have expanded their usage. Meanwhile, a few manufacturing clients churned, and some financial services deals never closed due to security concerns. In Phase 6, DataSync refines its ICP to “mid-size e-commerce companies on Shopify or similar platforms” as primary. Marketing shifts messaging to speak directly to e-commerce data pain points. Sales gets target account lists of top online retailers. The result: win rates climb and churn plummets over the next two quarters. By doubling down where the fit is strongest, they achieved more with less. That’s the power of ICP iteration.
Finally, note that ICP refinement is ongoing. Markets change – today’s ideal might face new pressures tomorrow. So periodically (say annually, or every couple of quarters) revisit: are the customers we thought were ideal still the best? Did a new segment emerge that we should test? Phase 6 is perpetual; you’re never really “done” refining ICP as you scale. But the heavy lifting of shifting your aim to the bull’s-eye happens now, as feedback floods in.

Iterating Messaging and Positioning – Tuning Your Value Proposition

If ICP is who you target, messaging is what you say to win them over. It’s your value proposition, your story, your differentiation articulated in words. Even great messaging at launch can become stale or mismatched once real customers engage with it. Phase 6 is when you take a hard look at your messaging through the lens of market feedback and performance data, and then fine-tune it for greater resonance.
Early on you might get by with a visionary story, but as you scale, messaging needs to be tested and optimized like any other part of the funnel. Here’s how to iterate your messaging:

Analyze conversion points for messaging effectiveness: Look at key points like email campaigns, ad click-through rates, website bounce rates, demo requests – these can signal if your message is connecting. For instance, if a specific email subject line or ad copy significantly outperforms another, dissect the language. Perhaps emphasizing a certain pain (“losing revenue due to manual errors?”) gets more response than a generic line (“improve efficiency”). Or maybe an eBook titled “A CFO’s Guide to Data Security” flopped, but a webinar “How SaaS Co. saved $1M with automation” filled seats, indicating perhaps your audience wants tangible ROI stories, not high-level guides. Use these clues.
**Run A/B tests deliberately: **different headlines, different landing page value propositions (one focusing on cost savings, another on speed, for example) and see which yields more sign-ups. Over a series of experiments, you’ll learn which messages resonate most with your ICP.
Gather qualitative feedback on messaging: This can be via formal means like message testing surveys or informally from sales and customer conversations. For example, have your sales reps ask prospects: “What part of our value proposition interests you most?” or “Did our description of the product make sense to you? How would you describe it back to me?” Listen for language: if the market consistently uses a different term for their problem than you do, consider adopting their language. Perhaps you’ve been saying “operational analytics platform” but customers keep calling it “reporting tool” – if so, maybe lean into more straightforward terminology. Also pay attention to confusion points: if prospects thought you did X when you actually do Y, your messaging might be misleading or too fuzzy. One technique: perform a message resonance survey with a sample of target buyers (there are services that show people a headline or a value prop statement and ask for feedback or understanding – some marketers use tools like Wynter for this). If half the respondents misunderstand what you offer, time to simplify and clarify the wording.
Revisit your core positioning statements: Dust off your elevator pitch, homepage headline, and other core messages. Do they still ring true? Do they incorporate what you’ve learned about customer priorities? For instance, maybe you initially positioned on “cutting-edge AI technology” but learned customers care less about the tech and more about “accurate results and time saved.” Then your messaging should shift to lead with outcomes (“Save 10 hours a week and eliminate errors”) rather than the how (“AI-driven engine”). Ensure your messaging hierarchy aligns with what actually resonates. A common framework is FAB – Features, Advantages, Benefits – but what really sticks with buyers are the benefits, especially if quantified. By Phase 6, you might have customer data to put into messaging (“Our clients reduce onboarding time by 50% on average”). Concrete claims beat abstract promises. If you have NPS or satisfaction results that are stellar, consider using them (“Rated 9/10 by finance teams for ease of use” – social proof).
Address common objections preemptively: Sales feedback is invaluable here. If you hear that in many late-stage deals the prospect was hung up on “Do you have X integration?” or “Is your security certified in Y?”, and your messaging never addresses those, you might incorporate that into materials. Not necessarily front-and-center if it’s secondary, but make sure your messaging isn’t silent on critical items. Alternatively, if a competitor is spreading FUD (fear, uncertainty, doubt) about your solution (“Their uptime isn’t reliable” or such), and you know it’s not true, bolster your messaging to counter that narrative (e.g., include your uptime stats or client trust logos). Iteration means adapting to the competitive and customer dialogue as it unfolds.
Test new angles, but keep a consistent identity: Iteration doesn’t mean changing your story every week – that confuses the market. Instead, it means gradually sharpening your core value prop. You might test a bold new tagline in a campaign, but you’re looking for evidence before rebranding everything. Once data indicates a better messaging angle, then roll it out broadly for consistency. For example, say you find “The all-in-one analytics hub for e-commerce” resonates way more than your current “Data platform for modern businesses” tagline, you might transition to the new messaging in your homepage, sales decks, LinkedIn page, etc. But ensure it still aligns with your fundamental vision. Don’t chase gimmicks; chase clarity and resonance.

One caution: when iterating messaging, ensure you bring your team along. Train sales on the new pitch, update collateral, and phase out old messaging in marketing materials to avoid confusion. For example, if Marketing starts emphasizing a new slogan but Sales is still using the old deck, prospects might get mixed signals. Alignment is key – perhaps include key messaging insights as part of your quarterly review outcomes.
Treat a major messaging revamp like a product experiment: did our key metrics move (higher conversion, shorter sales cycle, better survey responses)? If yes, great; if not, assess and iterate again. Message-market fit is a real thing: having the right message for the right ICP can dramatically accelerate GTM success. Phase 6 is where you ensure you achieve that fit.

Optimizing Channel and Tactic Mix – Doubling Down on What Works (and Fixing or Dropping What Doesn’t)

Your GTM strategy likely uses multiple channels: inbound marketing, outbound sales, partnerships, events, content marketing, maybe product-led growth loops, etc. As you measure performance, you’ll discover that some channels are far more effective for you than others. Phase 6 is about optimizing your channel mix – reallocating effort and budget to the winners, and improving or culling the underperformers. This is the continuous tuning of your GTM engine.
Let’s break down how to iterate on channels and tactics:

Evaluate channels with a metrics scorecard: For each major channel (or campaign type), review the cost, output, and conversion. For example:
- Content Marketing (SEO/blog): cost = mainly internal or agency content cost; output = website visitors and MQLs; conversion to SQL perhaps lower but long-tail.
- Paid Search Ads: cost = $X; yielded Y clicks, Z MQLs; CAC from this channel = maybe high or low?
- Webinars: cost (including time) vs leads generated vs how many became opportunities.
- Outbound SDR emails/calls: how many contacts, response rate, opportunities, wins.
- Channel Partners: how many referrals or deals from partners, and at what cost (e.g., commission or enablement effort).
Put it in a table for the quarter or year. You might see that, say, Webinars produced 200 MQLs at a cost of $50 each and 15 deals, whereas Paid LinkedIn Ads produced 500 MQLs at $200 each and 5 deals. That immediately tells you webinars are more bang for buck in this scenario. Or maybe inbound web leads close at a 10% rate, while cold outbound leads close at 2% – a huge difference in efficiency. Use CAC per channel if you can (at least roughly allocate spend by channel) and pipeline or ARR per channel. This quantitative view will highlight where to double down. One SaaS marketing team I recall found that one third of their spend (on a certain trade show) was driving almost zero ROI, so they cut it and reallocated to digital ads which were performing better.
Qualify the context: Raw numbers are great, but consider context. Some channels have longer-term benefits not captured in immediate metrics. For instance, content marketing (SEO) might have low lead yield this quarter, but it’s building an evergreen pipeline that lowers CAC over time. Brand campaigns (PR, thought leadership) might not show direct leads, but they raise conversion rates across all channels (people who saw your brand later click your ad more willingly). So, don’t kill a channel solely on short-term numbers if it plays a strategic role – but do set expectations and find other ways to measure impact (like brand awareness lifts for PR, etc. as discussed in qualitative metrics). On the flip side, if a channel is both expensive and not contributing to short or long term, it’s a clear cut for now.
Double Down on Winners: Once you identify a channel or tactic that’s working, consider scaling it up. For example, if webinars are delivering great leads, do more webinars. If one webinar topic did exceptionally well (say a panel on industry trends got triple the usual signups), consider a series on that theme. If your SDR outbound to a specific vertical is converting well, hire another SDR or invest in more contact data for that vertical. Kellogg’s funnel modeling approach emphasizes being able to scale spend where demand is strong – e.g., if you need more MQLs and webinars have proven capacity, throw more fuel on that fire quickly. But also ensure quality holds as you scale; sometimes a channel works up to a point, then saturates. Keep monitoring.
Fix or Drop the Losers: For channels underperforming, decide if the issue is fixable or if it’s fundamentally a bad fit. If it’s fixable, make a hypothesis and iterate. Example: “Our LinkedIn ads aren’t yielding SQLs. Maybe our targeting is off or the creatives are weak.” Try a change (new audience targeting, new ad copy highlighting a different pain point) for another cycle. If performance still lags, consider pausing or stopping that channel and reallocating budget. As a startup, your resources are precious; it’s better to focus on 2-3 channels that work than spread thin on 6 where 3 are wasted effort. However, maintain some experimentation budget to test new channels occasionally (maybe 10-20% of spend) because what doesn’t work now might work later or a new opportunity could emerge. Just be hypothesis-driven: e.g., “We’ve never tried a referral incentive program; let’s pilot it for a quarter with $10k budget and measure results.”
Channel-Strategy Alignment: Revisit if your channel mix matches your ICP and product. For instance, if you refined ICP to mid-market CIOs, are you reaching them in the right channels? Maybe you find a lot of effort was on Twitter and blogs, but those CIOs respond better to targeted executive dinners or LinkedIn outreach. Adjust channels to where your ICP “lives.” Similarly, if your product is complex and high-touch, a purely self-serve marketing approach might not cut it – you may need a sales-heavy (or event-heavy) approach. Or vice versa. Phase 6 might reveal these mismatches. A practical check: ask some recent customers, “How did you hear about us?” and “Where do you typically go to learn about solutions like ours?” Their answers might surprise you and indicate a channel you haven’t fully tapped (maybe a specific community or forum).
Account for Funnel Stages: Channel iteration isn’t only about lead generation. Consider the whole funnel: maybe Marketing is generating plenty of leads, but Sales struggles to convert them. Is it a channel issue? Possibly yes – leads from Channel X might be low quality tire-kickers. Or it might be a sales process issue. Examine conversion rates by channel as well. If one channel’s leads consistently convert to deals at a far lower rate, that channel might be feeding junk into the funnel (e.g., a content syndication program giving many names but poor fit). In such case, cut it, or significantly qualify those leads before passing to sales.
Also, consider post-sale when relevant: certain channels might yield customers who churn more. For example, perhaps those aggressive discount promotions on Black Friday brought many customers who left after the cheap period. Not great for SaaS. So channel strategy might include not pursuing certain tactics if they undermine customer quality.
Coordination and Sequencing: Over quarters, you may identify an optimal sequence or mix. For instance, some companies find success in an account-based marketing (ABM) approach: targeting key accounts with a combination of channels: ads + personalized outreach + executive events. If your review finds that multi-touch leads (those who engaged in 3+ programs) close at much higher rates, you might formally plan integrated campaigns rather than siloed channel efforts. This is iteration at the strategy level – realizing that channel synergy matters. On the flip side, you might find you’re double-counting or over-investing in overlapping channels (e.g., paying for leads on a directory site who are also coming through your organic funnel anyway). Then you’d streamline to avoid redundancy.

This iterative pruning and amplifying is like gardening: water the flowers (good channels) and pull the weeds (bad channels). But keep an eye that today’s weed might just be a late bloomer – so judge wisely with data over enough time.
One channel that often comes into play as companies mature is customer expansion and referrals – essentially turning your existing customer base into a “channel.” In Phase 6 you might realize that a lot of your growth is coming from happy customers expanding or telling others (if you measure things like the percentage of deals that came via referral). If so, that’s a channel to invest in: maybe formalize a referral program or put more resources in customer marketing. It ties back to NPS – a high NPS can be leveraged as a growth channel if you encourage referrals or case studies.
Iterating on your channel mix is about efficiency and effectiveness. It’s one of the fastest ways to improve GTM results without increasing budget – by reallocating the same spend/effort to higher-yield activities. It embodies Kellogg’s “growth efficiency” concept – focusing not just on growth, but growth at a reasonable cost. By Phase 6, you should be ruthlessly focusing on ROI of GTM tactics. The shotgun approach of early days evolves into a sniper approach: targeted, precise, and optimized. And just like every other aspect, keep measuring after changes. If you double spend on a channel, did it maintain performance at scale or hit diminishing returns? If a new experimental channel starts to show promise, nurture it. GTM strategy isn’t static because markets and buyers aren’t static. The companies that thrive are those that continuously refine their execution to find the optimal go-to-market fit for each stage of their journey.