08 · Competitive Analysis¶
Most competitive analysis produces a feature grid nobody reads. The reason is that it answers the wrong question. "What do they have that we don't?" generates a shopping list; the question you actually get asked in a roadmap review is "if we build this, will it still matter in a year, and if we don't, what do we lose?" Competitive analysis is how you answer that. It feeds four decisions you own: what to build, what to stop building, how to position, and what to charge. If your analysis doesn't change one of those four, you wrote a report instead of doing the work.
The other trap is defining the competitive set too narrowly. Your real competitor is usually not the other startup in your category — it's a spreadsheet, a contractor, or the customer deciding to do nothing.
Define the competitive set properly¶
List everyone who could absorb the budget and the job. Rank by how often they actually appear in your deals, not by how much they annoy you.
| Type | Definition | ListUp examples | Why it matters |
|---|---|---|---|
| Direct | Same job, same category, same buyer | Channelry and other multi-channel listing tools | Sets feature expectations and the price ceiling |
| Adjacent | Overlapping job, different centre of gravity | Marketplace-native bulk editors; storefront app stores | Can expand into you cheaply — watch their release notes |
| Substitute | Different mechanism, same outcome | Enterprise PIM at ~$600/mo; a VA paid to re-type listings | Often wins on trust, not features |
| Status quo | Doing nothing, or the spreadsheet | Sheets plus manual copy-paste | Usually your largest single competitor |
| Budget rival | Competes for the same money, different job | Ad spend, another marketplace subscription | Explains losses that otherwise look inexplicable |
The status-quo row is the one teams skip and the one that decides most outcomes. A seller who is "not currently evaluating anything" still has a working process. Beating it requires a switching argument, not a feature.
Where the evidence comes from¶
Rank sources by how close they sit to a real buying decision. Everything in the top block is worth a quarter's attention; the rest is context.
| Source | Tells you | Bias to correct for |
|---|---|---|
| Win/loss interviews | Why deals actually turned | Buyers give the polite reason first; ask twice |
| Churn interviews | What you failed to keep, and to whom | Recency — the last straw isn't the cause |
| Sales call recordings | Objections in the buyer's own words | Reps steer toward objections they can beat |
| Support tickets naming a rival | Migration friction and integration gaps | Only vocal users appear |
| Competitor pricing page | Packaging, value metric, target segment | Aspirational; list price ≠ paid price |
| Competitor changelog | Direction and investment pace | Marketing-filtered; silence ≠ inactivity |
| Their job postings | Where they're investing 6–12 months out | Slow signal, but hard to fake |
| Review sites | Recurring complaints, segment fit | Incentivised reviews skew positive |
| Analyst reports | Category framing and buyer language | Pay-to-play in some categories |
Buy the competitor's product. One month of a paid seat teaches more than a quarter of desk research, and it is the only way to see onboarding, error states, and the parts that are genuinely bad.
Compare on jobs, not features¶
A feature grid rewards whoever has the most checkmarks. A job grid rewards whoever does the thing customers actually hire the product for. Score each alternative on how well it does the job, not on whether it ships a feature named after it.
| Job to be done | ListUp | Channelry | Marketplace-native | Spreadsheet + VA |
|---|---|---|---|---|
| Publish one listing to many channels | Strong | Strong | None (one channel each) | Weak (manual) |
| Catch a pricing error before it goes live | Strong (diff preview) | Weak (post-hoc report) | None | Weak (human check) |
| Keep stock accurate across channels | Weak | Strong | Partial (own channel only) | Weak |
| Set up without a developer | Strong (median 41 min) | Weak (CSV mapping) | Strong | Strong |
| Handle 10,000+ SKUs | Weak | Strong | Strong | None |
| Prove to a client what changed | Partial | Weak | None | Strong (the sheet is the log) |
Read the columns, not the cells. ListUp wins where speed and safety matter and loses where scale and stock accuracy matter — which is a segment statement, not a backlog. The correct response to "Channelry is strong on stock accuracy" is first to ask whether our segment buys on it, and only then to consider building.
Turn findings into a decision¶
Every gap gets exactly one of four responses. Forcing the choice is the point — an unclassified gap silently becomes a roadmap item.
| Response | Use when | Cost | ListUp example |
|---|---|---|---|
| Differentiate | Already ahead, and the segment buys on it | Keep investing | Pre-publish diff preview — extend it to images |
| Neutralize | Table stakes; losing on it blocks deals | Cheapest credible version | Stock sync, good enough to stop being a reason to lose |
| Ignore | Real gap, wrong segment | Zero, plus a written rationale | 10,000+ SKU support |
| Concede | Structurally can't win; redirect the deal | Positioning work only | Enterprise procurement, SSO, custom SLAs |
Write the rationale down for the Ignore and Concede rows. Six months later someone will ask, and "we decided not to" is only credible with a date and a reason attached.
Win/loss analysis¶
Structured loss reasons are the highest-signal competitive input you have, because they are recorded at the moment money didn't change hands.
ListUp, Q3 — 380 trials that reached the evaluation stage:
| Outcome | Deals | % of evaluations | % of losses |
|---|---|---|---|
| Won | 152 | 40.0% | — |
| Lost to marketplace-native bulk tools | 91 | 23.9% | 39.9% |
| Lost to Channelry | 68 | 17.9% | 29.8% |
| Lost to spreadsheet + VA (status quo) | 49 | 12.9% | 21.5% |
| Lost to enterprise PIM | 20 | 5.3% | 8.8% |
Head-to-head win rate — the share of deals won when that alternative was named — is the more actionable cut, because competitors appear in very different volumes:
| Alternative | Deals where named | Won | Head-to-head win rate |
|---|---|---|---|
| Spreadsheet + VA | 118 | 69 | 58.5% |
| Channelry | 143 | 75 | 52.4% |
| Enterprise PIM | 42 | 22 | 52.4% |
| Marketplace-native tools | 167 | 76 | 45.5% |
Two findings fall out. The biggest loss bucket is marketplace-native tools, where the head-to-head rate is also the worst — those are sellers on one or two channels who don't yet need cross-listing, so this is a targeting problem, not a product one. Second, splitting the Channelry deals by segment is decisive:
| Segment | Channelry deals | Won | Win rate |
|---|---|---|---|
| Sellers on 1–2 channels | 62 | 21 | 33.9% |
| Sellers on 4+ channels | 81 | 54 | 66.7% |
Same competitor, same product, nearly double the win rate in one segment. That one split is worth more than any feature comparison: it tells marketing who to target, sales which deals to work, and the PM that the roadmap gap is real only in the segment we already lose.
The battlecard¶
One page, written for someone on a call with twenty seconds to read it.
| Section | Content | Length |
|---|---|---|
| Who they are | One sentence, plus their real target segment | 1 line |
| When we win | The two situations where we're clearly better | 2 bullets |
| When we lose | Honest — reps only trust a card that admits this | 2 bullets |
| Their strongest claim | Stated fairly, then answered | 1 + 1 |
| Landmine questions | Questions that expose their weakness without naming them | 3 |
| Never say | Claims that are false, unprovable, or invite a demo we lose | 2–3 |
Landmine questions beat direct attacks. "How do you handle a listing that exists on three channels with different attribute names?" does more work than "Channelry is bad at mapping" — and it doesn't age badly the week they ship a fix.
Keeping it alive¶
| Cadence | Activity | Owner |
|---|---|---|
| Continuous | Log a loss reason on every closed deal (mandatory field) | Sales |
| Monthly | Read win/loss; update head-to-head rates | PM |
| Quarterly | Refresh the job grid; re-run the four-response classification | PM |
| Quarterly | Re-buy and actually use the top competitor's product | PM |
| On trigger | Rival raises funding, ships a major release, or changes pricing | PM, within 1 week |
An analysis refreshed annually is a historical document. The mandatory loss-reason field is the cheapest item on this list and the one that makes everything else possible.
Worked example — ListUp's Q3 competitive review¶
Trigger: Channelry announced funding and launched a free tier.
What the data said: the free tier caps users at two channels, which targets exactly the segment where our head-to-head win rate is 33.9% and where we were already losing to marketplace-native tools. Among 4+ channel sellers we win 66.7%, and the cap excludes them entirely.
Decisions taken:
- Concede the 1–2 channel segment as an acquisition target. Stop spending on the keywords that bring them in — they convert poorly and churn fast.
- Neutralize stock accuracy. It was the top named reason in 29 of the 68 Channelry losses. Build the cheapest credible sync, not the best one.
- Differentiate harder on the diff preview, which appears in no competitor changelog and maps to the pricing-error pain already quantified at roughly $40 per incident.
- Ignore 10,000+ SKU support. Rationale filed: the median account has 340 SKUs and the 95th percentile is 2,100.
Why this survived the roadmap review: no line item came from "Channelry has it." Each came from a loss count, a segment win rate, or a written decision not to act — including two decisions to build nothing, which is the part most competitive reviews are missing.
How It Actually Works¶
Competitive analysis works by triangulating a competitor's revealed strategy (what they're actually building and charging, visible in the market) against their likely constraints (funding stage, team size, technical architecture) to predict their next move — the mechanism is that most competitive moves are highly constrained by what a company can plausibly execute given its resources, so a well-informed guess about their constraints narrows the space of plausible next moves far more than guessing based on stated intentions (which are cheap to say and unreliable). Feature-parity chasing is a common trap explainable by a specific mechanism: matching a competitor's feature list treats their public surface as a proxy for their strategy, but features are the output of a strategy shaped by their specific customer base and unit economics, so copying the output without the underlying strategic context often produces a feature that serves your competitor's customers' needs, not yours. Win/loss analysis works because it's the only source of causal competitive data (why a specific real deal was actually won or lost) as opposed to inferred competitive data (what you assume based on feature comparison) — the mechanism that makes it valuable is that it samples from the actual decision process of real buyers under real constraints, which is a far more reliable signal than any amount of desk research about a competitor's public positioning.
Exercise¶
Run a real competitive review for your product and produce:
- A competitive set table covering all five types, including status quo. For each, state how often it appeared in your last 20 closed deals.
- An evidence plan: three sources, what each can and can't tell you, and the bias you'll correct for. Include one source that requires actually using a competitor's product.
- A job-based comparison grid: 5–6 jobs, 3+ alternatives, scored on outcome quality rather than feature presence. Then write the one-sentence segment statement the columns imply.
- A four-response classification of every gap you found, with a written rationale for each Ignore and Concede.
- A win/loss table for last quarter, with both loss-share and head-to-head win rates, plus one segment split. If you don't record loss reasons today, your first deliverable is the field, not the analysis.
- A one-page battlecard for your top competitor, including the "when we lose" and "never say" sections. Show it to a salesperson and cut anything they wouldn't use on a live call.