How to shortlist startups for a pitch event: a rubric

Shortlisting is where most pitch events are won or lost. Pick on polish and you get a stage of good presenters. Pick on evidence and you get startups investors want to meet.

This guide gives a weighted rubric, then covers the three things that make it fair in practice: conflicts of interest, reviewer calibration and handling volume. It builds on our note on startup evaluation criteria for incubators, which explains where pitch-led scoring goes wrong.

What should a pitch event shortlisting rubric score?

A shortlisting rubric should score five things on a 10-point scale: evidence of demand, product built, team, problem and market clarity, and fit with the event. Evidence-based items carry 7 of the 10 points. Presentation quality carries none, because the pitch itself will test it.

Criterion	Points	Full points for	Earns nothing
Evidence of demand	3	Paying customers, signed pilots, repeat usage	"Huge market", waitlist sign-ups alone
Product built	2	A working product with real users	Mock-ups, a roadmap
Team	2	Founders who have shipped, full-time	Advisor logos
Problem and market	2	A named buyer and a reason to buy now	Top-down market size
Fit with this event	1	Stage and sector suit the investors attending	A strong startup for a different room

Score the stage, not an absolute bar. An idea-stage startup with ten customer interviews and a prototype can earn full points for demand at an idea-stage event. Write that rule into the rubric before reviewers start.

Ask reviewers to note the evidence behind each score in a few words, such as "two paying clients, invoices mentioned" or "pilot unpaid". A score with a reason can be checked in the shortlist call. A bare number can't, and it hides the reviewer who scored on gut feel.

The rubric only works if the form collects the evidence. The pitch event application form template asks for deck, team, stage and revenue; add one question such as "What is your best proof that customers want this?"

Why score evidence over polish?

Polish measures how well a founder presents, which the pitch itself will show. Evidence measures what the startup has done, which a deck can hide. Scoring evidence at shortlist stage stops the best-designed decks from crowding out the startups with the most progress.

Official rubrics lean the same way. The sixth National Startup Awards' student category proposes 50 points for innovation and problem-solving, 30 for scalability and 20 for economic impact. Nothing goes to presentation.

Smart India Hackathon's 2026 guidelines list novelty, complexity, clarity, feasibility, practicability, sustainability, scale of impact and user experience. Clarity appears, but as one item among eight.

Watch for the deck habits that score well and mean little. Our list of red flags in startup pitch decks covers the common ones.

Score what the startup has done, not how well the deck says it.

How do you handle conflicts of interest?

Collect conflict declarations from every reviewer before scoring starts, and reassign any application where a reviewer has a link to the startup. Ask applicants to declare known conflicts too. Two-sided declarations catch what one side forgets, and a written log protects you when a result is questioned.

A reviewer is conflicted if they, or their firm, have:

• invested in or been offered equity in the startup;
• an advisory, mentoring or employment role with it;
• a family or close personal link to a founder;
• a competing company or a pending deal in the same space.

The rules for the sixth National Startup Awards make applicants disclose actual or potential conflicts involving evaluators, jury or committee members. A pitch event can copy that line into its application form at no cost to anyone.

Mentors from your own incubator are the most common conflict. They know the startups best, so let them brief the jury on facts, but keep them off the score sheet for their own cohort.

How do you calibrate reviewers?

Before real scoring, have every reviewer score the same small set of applications independently, then compare. Where scores differ by 3 points or more out of 10, discuss why and write down an example of what earns each score. Then score the rest, two reviewers per application.

1. Pick a sample. Five applications: a clear yes, a clear no and three in between.
2. Score blind. Nobody sees another reviewer's marks until all are in.
3. Discuss the gaps. Agree what "3 of 3 for demand" looks like at this stage.
4. Write anchors. One line per criterion, with an example, on the score sheet.
5. Resolve splits. A third reviewer scores any application where two totals differ by 3 or more.

A common score sheet across reviewers is what makes this work. The National Startup Awards' student category asks every State and UT to evaluate on a common score sheet for the same reason: scores from different panels must mean the same thing.

How do you handle hundreds of applications?

Handle volume with stages and caps. A yes or no eligibility screen comes first and needs no scoring. Two scored reviews follow for the applications that pass. Caps on entries, set in advance, stop the queue from outgrowing your reviewers.

The large government programmes cap at every stage. Smart India Hackathon 2026 accepts only 500 ideas per problem statement, after which that statement closes. Each college may nominate up to 50 teams, 45 shortlisted and 5 waitlisted, and about 4 to 5 teams per problem statement may reach the finale. The National Startup Awards' student funnel takes about 360 nominations down to eight awardees.

For a pitch event, the same logic gives three rules. Cap applications or close early if reviewer time is short. Keep a waitlist of two or three startups for dropouts. And split review across sectors, so each reviewer sees startups they can judge.

Give each reviewer a batch and a deadline, not an open queue. Ask them to score the batch in one or two sittings, so their standard doesn't drift over weeks, and the deadline tells you early when a reviewer has dropped out. Rank the scored list once, at the end, rather than accepting startups as scores arrive.

Angel networks face the same volume and solve it in stages; see how angel networks review startup applications. Once the list is final, tell every applicant on the date you promised, as set out in our guide to the demo day application process.

Questions

What criteria should you use to shortlist startups for a pitch event?

Score evidence of demand, a working product, the team's record of shipping, clarity about the problem and market, and fit with the investors attending. Weight evidence highest. A deck's design, logos and buzzwords should earn nothing, because they measure presentation skill rather than progress.

How many reviewers should score each pitch application?

Use at least two reviewers per application, scoring independently without seeing each other's marks. When their totals differ by 3 points or more out of 10, send the application to a third reviewer. Calibrate everyone on the same sample of applications before real scoring begins.

How do you handle conflicts of interest when shortlisting startups?

Ask every reviewer to declare investments, advisory roles, employment, family links and competing interests before scoring, and reassign any conflicted application. Ask applicants to declare known conflicts too. The sixth National Startup Awards require applicants to disclose conflicts involving evaluators, jury and committee members.

Sources: Startup India, DPIIT, NSA 6.0 brief note on Student Innovator category, Oct 2026 · Startup India, DPIIT, Guidelines for National Startup Awards 6.0, Oct 2026 · Smart India Hackathon, SIH 2026 Guidelines (checked 10 Nov 2026)

Build your event form on MyStartupValley Events: mystartupvalley.com/msvevents. More organiser guides are on the MyStartupValley blog.