How to Read Web Development and Ad Campaign Estimates: Man-Month Pitfalls and Validity Checks
Published:
Last Updated:
Published:
Last Updated:
Category: Marketing Budget & KPI, Marketing Strategy
Authors: Shusaku Yosa
When you outsource web development or an advertising campaign, it is common to line up quotes from several vendors and still have no idea which one is reasonable. One quote is twice the price of another, yet both appear to describe the same website. That uneasy feeling almost always traces back to differences in assumptions and in how effort was built up. This article covers how to read estimates for web and advertising work, the pitfalls hidden in man-month pricing, and the checks that tell you whether a quote holds up.
The main reason is that the deliverable is not a standardized product. Two projects described identically as a ten-page corporate site can require several times the effort depending on whether the design is created from scratch or adapted from a template, who supplies the copy and photography, whether a CMS is involved, and how far the mobile experience is built out.
Estimate formats also vary by company, which compounds the problem. Some vendors break work down into fine line items and build the total upward; others write a single line reading "site production, lump sum." The first looks expensive and the second looks cheap, even when the actual workload is identical.
Comparing estimates, in other words, is not the work of comparing numbers. It is the work of aligning assumptions. Skip that step and a change order will surface after you sign, in the form of "that task was not included."
Read these five places first. An estimate that leaves them blank is not yet something you can compare.
Check whether the work is broken out by phase: requirements, information architecture, design, implementation, testing, launch. If it is, you can see where the cost concentrates and judge what can and cannot be trimmed. A single line reading "production, lump sum" leads straight to the first pitfall described below.
"Design, 1 lot, 800,000 yen" tells you nothing. "Design: 1 homepage plus 3 subpage templates, X person-days at Y yen" lets you trace what you are paying for. The unit of quantity matters too. Per-page pricing is easy to understand but tends to hide how much more work some pages require than others.
The assumptions printed at the bottom of the estimate are where the price difference is actually explained. Lines such as "copy and images supplied by client," "browser support limited to the latest two versions," "two rounds of design revision," or "server and domain contracted by client" are the boundaries of the scope. A cheaper quote usually has stricter boundaries here.
What is excluded matters as much as what is included. Copywriting, photography, multilingual versions, post-launch maintenance, analytics setup, and data migration from an old site are all commonly assumed to be included when they are in fact billed separately.
Quotes are typically valid for one to three months. If internal approval takes time, plan for the possibility of a re-quote. Payment terms matter to cash flow as well: half on kickoff and half on delivery is very different from a single payment after acceptance. Where payment is split, confirm what triggers each installment, whether that is kickoff, an interim deliverable, or final acceptance.
Most web and software estimates are assembled as monthly rate multiplied by effort. One person-month means the volume of work one person completes in a month of full-time effort, treated in practice as roughly 20 person-days, or about 160 hours at eight hours per day.
So a staffing plan of 0.5 person-months of design, 1.0 of front-end engineering, and 0.3 of project direction produces a base figure by multiplying each role by its monthly rate and adding them up. Any gap between two quotes therefore comes from the rates, from the effort estimate, or from both.
The ranges below are what you commonly see at Japanese web and development firms. They shift with company size, location, seniority, and contract type, so treat them as rough reference points.
What this tells you is that an unusually cheap quote is likely either staffed with lower-rate people or built on an effort estimate that is thinner than the work requires. Figuring out which one it is becomes the next step in your comparison.
Knowing which band your project falls into gives you a first-pass judgment on whether a quote sits at a sensible level. When a number lands well outside the range, high or low, always ask why.
"Production, lump sum, 3,000,000 yen" gives you nothing to judge against. You cannot negotiate, and you cannot trim scope to fit a budget. When you see a lump sum, ask for a phase-level breakdown with estimated effort, even informally. A vendor unable to produce one may have priced the job by feel rather than by building it up.
Vendor A charges 800,000 yen per person-month and vendor B charges 1,200,000. On rate alone A wins. But if A estimates three person-months and B estimates 1.5, the totals are 2,400,000 against 1,800,000 and B is cheaper. Stronger practitioners deliver the same output in less time, so rate and total do not necessarily move together. Compare totals against deliverables, not rates.
Treating one person-month as 20 person-days assumes that person spends every working day on your project alone. In reality, parallel projects, meetings, waiting on reviews, and environment troubles eat into it. Schedules slipping while the effort numbers look correct usually traces back here. Asking how many people at what utilization will absorb the estimated effort tells you how realistic the timeline is.
Progress tracking, internal review, meetings, minutes, and specification adjustments produce no visible deliverable and are often left out. Direction and management commonly account for something like 10 to 20 percent of production cost. A quote showing zero here looks cheaper, but the likely outcome is either thin oversight or a later invoice.
The most frequent source of cost disputes is the gap between what the client considers a minor tweak and what the vendor sees as redoing the design. At estimate stage, confirm how many revision rounds are included, where the line falls between a revision and a specification change, and what process applies when a change occurs, such as issuing a change request and agreeing on it before work begins. With a rule in place, additional cost becomes an expected decision rather than an ambush.
You live with a site far longer than you build it. Hosting and domain fees, SSL, CMS and plugin updates, incident response, minor content updates, and analytics reporting all continue afterward. Comparing initial cost plus three years of running cost sometimes reverses which vendor looks better.
Work through these eight points in order and you will get most of the way to a judgment.
Media buying quotes are structured differently from production quotes. Because the core charge is a fee tied to media spend rather than to effort, the things worth checking change as well.
Of the total you pay, confirm exactly how much reaches the ad platforms and how much is the agency's management fee. In Japan, a fee of 20 percent of media spend is a widely used convention, though some agencies work on a flat monthly fee or a performance-based model. When a quote simply reads "1,000,000 yen per month," always ask what share of that lands on the media.
Most agencies set a monthly minimum, often in the range of 50,000 to 100,000 yen. While media spend is small, that floor rather than the percentage governs, so the effective fee ratio on a small budget is higher than it appears. At the other end, when media spend is large, it is fair to test whether 20 percent still matches the actual work involved. Asking how many hours and how many people the account will receive each month lets you reframe the fee in person-month terms.
Whether banner and ad copy production sits inside the fee or is billed separately varies by agency. Even when included, there is usually a cap such as a set number of assets per month. Because refresh frequency drives performance in performance advertising, the monthly asset count and the unit price for anything beyond it are worth pinning down before you sign.
Monthly reporting only, or weekly updates? Regular meetings in person or online? This drives a large share of the agency's workload and is a useful gauge of whether the fee is fair. The content matters too: a report that lists numbers is worth less than one that carries recommendations for the following month. Asking for a sample report in advance is the reliable way to find out.
Anything explained verbally will be interpreted differently by each vendor. Put the objective, scope, page count, required functionality, what you supply, target launch date, and budget range on a single sheet, and hand the identical document to everyone. That step alone dramatically improves comparability.
Asking vendors to break the quote down by phase with estimated person-days, and to separate initial from running costs, brings every submission to the same level of detail. Specifying the format is not rude; it is a legitimate requirement for comparison.
Where you find a large gap on a line item, question both the high and the low side. Why is this phase weighted so heavily? What makes it possible to deliver this in less effort than others quoted? The answers reveal how well each vendor understands the problem, and how straight they are with you.
Understanding of the problem, track record, team, ongoing support, and ease of communication. Weighting these before you open the quotes keeps price from dominating the decision. Choosing on price and then paying it back in rework and management overhead is a failure this one step prevents.
Not necessarily. Heavier investment in requirements and design, more thorough testing, or copywriting included in scope all routinely explain a gap of that size. Start by comparing the stated assumptions side by side, and if the difference still is not explained, ask for the reasoning behind the breakdown. A quote whose logic cannot be explained is hard to trust regardless of the number.
Share the range anyway. Withhold it and you get a mix of impossibly cheap proposals and clearly over-budget ones, which makes comparison impossible. Framing the request as asking for the strongest plan achievable within the range tends to prevent proposals that simply consume the ceiling, and how each vendor prioritizes reveals a great deal about their capability.
Most often from page count growth, a reversal on design direction, schedule slippage caused by late delivery of copy and images, specification additions just before launch, and internal requests arriving late. Many of these originate on the client side. Consolidating internal requirements up front and confirming who has final sign-off is, in the end, the most effective cost control available.
Around three is realistic. Preparing an estimate costs the vendor real time, and every additional company adds to your own comparison workload. Narrowing the field first on track record and specialization raises the quality of what each one submits.
Read web and advertising estimates through three lenses rather than the total: the breakdown, the stated assumptions, and what is out of scope. Once you understand how man-month pricing is assembled, you can separate a rate difference from an effort difference and shift the conversation from haggling to adjusting scope, which is a far more productive place to be.
Lump-sum lines, comparing on rate alone, optimistic utilization, missing project management, no change-control rule, and overlooked post-launch cost. Avoiding these six pitfalls alone will remove most of the trouble that follows a signature. Start by opening the quote in front of you and reading the assumptions first.
How to write a requirements definition document, with a seven-chapter template you can use as-is. Covers turning vague w...
A creative brief tells agencies and production studios your background, objective, and constraints. How it differs from ...
What vendor management means, why a hands-off approach fails, the four areas to manage, five things to settle before eng...
