EarlBear growth

Life without EarlBear: pricing the growth team you would have to hire

By Omar · · 17 min read

Say your store works. Traffic is real, orders arrive daily, and revenue has found a rhythm — call it 50,000 sessions a month converting at 2% on an $80 average order, just under $1M a year.1 Then growth flattens, and the question becomes the one every founder eventually types into a search bar: how do I take this to the next level?

The honest answer, without software that does it for you, is: you hire a team. Growth is not a tactic you bolt on; it is a full-time loop of analysis, hypothesis, prioritization, building, testing, and measurement, and every step of that loop is someone’s job. This post prices that answer — the roles, the salary bands, the calendar, and the monthly bill — so you can see what “do it manually” actually costs before you choose it.

What growth work actually is

Growing a store that already has traction is not one big move. It is a queue of small, testable changes, each of which has to earn its place:

  1. Analyze — find where the funnel leaks (product page, cart, checkout, post-purchase).
  2. Hypothesize — propose a change that should close the leak.
  3. Prioritize — rank it against everything else by expected impact and effort.
  4. Build — design, implement, and QA the change behind a flag.
  5. Test — run an A/B test until the result is statistically conclusive.
  6. Decide — ship the winner, revert the loser, log what you learned.

Each pass through the loop takes real calendar time, and the build step alone varies by an order of magnitude:

ChangeBuild time
Copy, pricing display, or CTA tweak1–2 days
New landing page or PDP layout1–2 weeks
Site speed and Core Web Vitals work1–3 weeks
Checkout flow change2–4 weeks
Recommendations or merchandising logic3–6 weeks

And building the change is the fast part. The slow part is finding out whether it worked.

The A/B test math nobody warns you about

An A/B test needs enough visitors to tell signal from noise. The standard approximation for the sample size per variant (80% power, 5% significance) is 16·p(1−p)/δ², where p is your baseline conversion rate and δ is the absolute lift you want to detect.2

Plug in the store above1 — 2% baseline conversion:

  • Detecting a 10% relative lift (2.0% → 2.2%) needs about 78,000 visitors per variant — roughly 157,000 total. At 50,000 sessions a month, that is more than three months for one conclusive test.3
  • Detecting a 20% relative lift (2.0% → 2.4%) needs about 20,000 per variant — call it 3–4 weeks, if you send it all your traffic.3

Add one to two weeks up front for instrumenting the test — variant assignment, event tracking, QA on both arms — and the practical cadence at this traffic level is roughly one conclusive experiment per month, and only for changes big enough to plausibly move conversion 20%.3 Subtle ideas are untestable at this scale; you simply cannot afford the calendar.

That cadence is the constraint everything else hangs on: if you can only get about twelve real answers a year, the people picking and building those twelve bets had better be good. Which is why the hiring bar — and the payroll — looks the way it does.

The team you would hire

A minimal in-house growth team for a store at this stage is six people. The salary bands below are 2026 US base-salary figures, drawn from published compensation guides for each role: growth lead,45 senior product manager,6 senior frontend engineer,7 senior backend engineer,8 growth data analyst,9 and support / operations engineer.10 Your market and remote policy will move them, which is what the calculator below is for.

Base salary band by role, with the median marked
Growth lead8+ yrs; has owned a DTC or marketplace P&L
$150–220k
Senior product manager5+ yrs; experimentation background
$130–190k
Senior frontend engineer5+ yrs; storefront performance, feature flags
$130–185k
Senior backend engineer5+ yrs; integrations, checkout and pricing
$135–195k
Growth data analyst3–5 yrs; experiment statistics, cohort analysis
$110–145k
Support / operations engineer4+ yrs; SRE background; on-call, incident response
$130–175k

The sixth seat is the one founders forget until the first outage. Every change the team ships is another thing that can break at 2 a.m. — a checkout regression, a third-party integration that starts timing out, a deploy that tanks Core Web Vitals. The support / operations engineer owns the unglamorous half of growth: monitoring and alerting, an on-call rotation, incident response and postmortems, and the SLOs that turn “the site feels slow” into a number someone is accountable for. Without that seat the experiment velocity you paid for leaks straight back out as downtime and firefighting.

Who owns what

Hiring the six roles is not the same as knowing who does what when a change moves through the loop. This is where the manual path gets expensive in coordination, not just salary: every step needs one accountable owner and the right people consulted, or decisions stall. Mapped as RACI — Responsible (does the work), Accountable (owns the outcome), Consulted (weighs in), Informed (kept in the loop):

Loop stepGrowth
lead		Product
mgr		Front­endBack­endAnalystSupport
/ops
Analyze the funnelACIIRI
HypothesizeCA·RCCCI
PrioritizeARCCCC
Build the variantIARRIC
Run the A/B testIARCRI
Decide and shipARCCCC
Operate and keep it healthyICCCIA·R

The pattern that matters: the support / operations engineer is Accountable for the one row nobody else wants — keeping what you shipped alive — and Consulted on every build and ship decision, because reliability is cheapest when it is designed in, not bolted on after the incident.

Three multipliers sit on top of base salary:

  • Loaded cost. Payroll taxes, benefits, equipment, and tooling add 25–40% on top of base.11 We use 1.30 as the default below.
  • Recruiting. Filling a role takes about 44 days on average — senior searches run longer12 — and a contingency agency costs 20–25% of first-year salary as a one-time fee, per role.13
  • The top-of-market premium. The bands above are mid-market. If you want the people who are demonstrably the best — the top decile, the ones who make the whole loop faster — you pay for the tail of the distribution, not the middle. Across engineering roles, 90th-percentile (“top of market”) base pay runs roughly 60–85% above the median, and even reliably landing senior standouts usually means 15–20% over the going rate.14 The calculator’s talent tier slider lets you dial from median to top-of-market and watch the bill move.

The calendar: nine months to a working team

You cannot hire all six at once, because the first hire chooses the rest. The growth lead takes two to three months to land (search, interviews, notice period), ramps for a month, and only then starts hiring the product manager, engineers, analyst, and support engineer. Here is a realistic sequence:

Time to a productive growth team, by role
Growth lead
Product manager
Frontend engineer
Backend engineer
Growth analyst
Support engineer

The first experiment goes live around month six. Given the test math above, the first statistically conclusive result — the first time you know a change worked — lands around month eight or nine. Until then you are paying full payroll for setup.

The math, on your numbers

Every store is different, so here is the whole model in one widget: toggle roles, adjust salaries to your market, set the overhead multiplier and your realistic experiment cadence, and read off the bill. Defaults are the mid-band figures from the table above.

Monthly team cost–
Annual run rate–
First-year total–incl. 20% recruiting fees
Cost per conclusive experiment–

With the default numbers — all six roles at mid-band, a 1.30 overhead multiplier — the team costs about $102,000 a month,3 about $1.23M a year in run rate, and roughly $1.41M in year one once recruiting fees are in. At one conclusive experiment a month, every real answer about your store costs about $102,000 — and the first one arrives eight to nine months after you start the search for your growth lead.

None of this is an argument that the people are overpaid. It is what skilled people cost, doing a loop that genuinely requires skill. It is an argument about the shape of the work: for a store under a few million in revenue, the manual path spends well over a million dollars a year to run about twelve experiments — and to keep the results alive once they ship.

Why we are building EarlBear

That loop — observe the funnel, form a hypothesis, build the variant, run the test, ship the winner, watch for regressions — is exactly the loop EarlBear runs as software. Self-healing ecommerce means the analysis never sleeps, the test queue never waits on a hiring pipeline, and the cost scales with your store instead of with US salary bands. Life without EarlBear is not impossible; it is just nine months and a million dollars away.

Footnotes

  1. The store in this post is illustrative — 50,000 sessions a month, a 2% conversion rate, and an $80 average order value, chosen to sit near typical mid-market ecommerce figures. Substitute your own numbers in the calculator. ↩ ↩2

  2. Miller, Evan. “Sample Size Calculator.” Evan’s Awesome A/B Tools, n.d., https://www.evanmiller.org/ab-testing/sample-size.html. ↩

  3. Derived figure — computed from the scenario assumptions, the cited salary bands, and the sample-size formula, not taken from an external source. The calculator recomputes it from your inputs. ↩ ↩2 ↩3 ↩4

  4. “eCommerce Salary Guide 2026: Compensation Benchmarks for Every Role.” eCommerce Placement, 2026, https://www.ecommerceplacement.com/resources/ecommerce-salary-guide-2026/. ↩

  5. “Director of Growth Salary in the United States.” Salary.com, 2026, https://www.salary.com/research/salary/listing/director-of-growth-salary. ↩

  6. “The Hard Truth About Product Management Salaries in 2026.” Product School, 2026, https://productschool.com/blog/career-development/product-management-salaries-todays-economy. ↩

  7. “2026 Software Engineer Salary in US.” Built In, 2026, https://builtin.com/salaries/us/software-engineer. ↩

  8. “2026 Salary Guide: Software Engineers and Developers.” Motion Recruitment, 2026, https://motionrecruitment.com/it-salary/software. ↩

  9. “Data Analyst Salary Guide 2026: Pay by Level, City & Skill.” KORE1, 2026, https://www.kore1.com/data-analyst-salary-guide/. ↩

  10. “Site Reliability Engineer Salary Guide 2026.” KORE1, 2026, https://www.kore1.com/sre-salary-guide-2026/. ↩

  11. Hadzima, Joseph. “How Much Does an Employee Cost?” MIT Sloan School of Management, Boston Business Journal series, n.d., https://web.mit.edu/e-club/Archive/hadzima/pdf/how-much-does-an-employee-cost.pdf. ↩

  12. “Cost Per Hire Calculator: SHRM Formula & Benchmarks.” Teamed, 2026, https://www.teamed.global/insights/cost-per-hire-calculator-shrm-formula-and-benchmarks. ↩

  13. “How Much Do Recruitment Agencies Charge?” Leonar, 2026, https://www.leonar.app/blog/how-much-do-recruitment-agencies-charge/. ↩

  14. “Engineering: Average Salary & Pay Trends 2026.” Glassdoor, 2026, https://www.glassdoor.com/Salaries/engineering-salary-SRCH_KO0,11.htm. The 60–85% figure is the gap between median and 90th-percentile base pay across engineering roles reported here and in the U.S. Bureau of Labor Statistics percentile wage data (“Percentile Wages.” U.S. Bureau of Labor Statistics, 2024, https://www.bls.gov/oes/oes_perc.htm). ↩