From Rankings to Revenue: Building an SEO Business Case with Semalt Data

Every SEO engagement eventually reaches the meeting where someone outside marketing asks what the money bought. It is a fair question, and the industry answers it badly — usually with a chart of positions, which is a description of the method rather than the result.

This article covers how to build a defensible business case from the data in the Semalt platform: valuing a keyword set before the work starts, forecasting in a way that survives contact with reality, and reporting returns without inflating them. It is written for the conversation with a finance director, not for the conversation with a marketing manager who already believes in search.

Valuing a keyword set before you start

The calculation is not complicated, and its value is that it forces every assumption into the open where it can be argued with.

1. Start with the commercial subset only

   Take the keywords with buying intent. Exclude branded terms — you already have that demand — and exclude informational terms from this particular calculation, because their contribution is real but indirect and mixing them makes the model dishonest.

2. Apply a realistic click share

   Not "we will be first". Model a target position band and use a conservative click-through estimate for that band, then reduce it further for queries where the results page carries ads, a map pack or an answer box above the classic results.

3. Use the client's own conversion rate

   Not an industry benchmark. If they do not know it, that is the first thing to fix, and the fact that they do not know it is itself worth raising.

4. Multiply by margin, not by revenue

   A shop with a twelve per cent margin and a consultancy with a seventy per cent margin can support very different SEO budgets on identical traffic. Finance people notice immediately when you use revenue, and it damages the credibility of everything else in the document.

Structure of the model described here; the numbers themselves must come from the client's own data.

Forecasting without lying

The temptation in a proposal is to produce a confident number, because confident numbers win pitches. They also produce the meeting nine months later where you explain why the confident number did not happen.

The professional alternative is a range with visible assumptions. Three scenarios, each stating what has to be true for it to hold: a conservative case where the client's development team is slow and competitors continue investing, a base case matching current conditions, and an upside case that requires specific things to go right. When one of those assumptions later fails — the developer never shipped the fixes, a competitor was acquired and started spending — you have a documented reason rather than an excuse.

What SEO can and cannot claim

Attribution honesty is the fastest route to being trusted in a room full of people who have heard marketing claims before.

The equivalent-paid-cost figure in the first column is the single most effective number we use with finance stakeholders. Take the non-branded organic sessions the tracked set produced, apply the cost-per-click those same terms carry in paid search, and you get a defensible figure for what renting that traffic would have cost this month. It is not the same as profit, and you should say so — but it converts an abstraction into a comparison with a budget line the reader already understands.

Two arguments that will be made against you

Being ready for them is part of the business case.

"Those customers would have found us anyway." Sometimes true, particularly for branded search. This is precisely why the model excludes branded terms and reports non-branded separately. If the growth is happening on commercial terms the business did not previously rank for, the counterfactual is weak and you can show it.

"Paid search gives us the same thing faster." True on the timescale, false on the arithmetic once the campaign has run for a while. Paid traffic stops the day the budget stops; organic positions decay slowly. The honest framing is that they are different instruments — one is rented reach with immediate delivery, the other is an asset with a build period — and most businesses need both. Arguing that SEO replaces paid search is a losing position and an untrue one.

See also: From the Greek Market to International.

Reporting returns month to month

Once the engagement is running, the reporting has to connect back to the model or the original business case quietly becomes irrelevant.

Note the order. Most SEO reports are built bottom-up, opening with diagnostics and reaching outcomes on page eleven if at all. Inverting that single decision changes how the work is perceived more than any improvement in the underlying data.

The measurement gap that undervalues most local businesses

In a large share of service businesses, the conversion is a phone call rather than a form submission. If calls are not measured, the business case systematically understates the channel — frequently by most of its actual value — and you will be arguing for renewal with a number you know is wrong.

Click-to-call tracking on mobile is the minimum, and it is trivial to implement. For businesses with meaningful call volume, proper source attribution is worth the setup: without it, every enquiry that arrives by phone is credited to nobody, which in practice means credited to whoever reports most confidently.

Raise this in the first month rather than the ninth. A client who agrees to measure calls at the start gets an honest picture throughout; a client asked to believe in unmeasured calls after nine months hears an excuse, however true it happens to be.

We cover this in detail in The First 30 Days with Semalt.

What this means for how you price

A defensible model changes the commercial conversation in a way that benefits both sides. If the tracked commercial set can realistically support a specific level of monthly margin at a target visibility, then a retainer priced well below that has an obvious justification, and one priced above it does not deserve to win.

It also identifies the clients you should not take. If the arithmetic says the entire achievable commercial demand in a niche is worth less per month than your fee, no amount of skill fixes that, and taking the engagement guarantees an unhappy ending. Saying this out loud in a first meeting costs you a project occasionally and earns you a reputation that returns much more.

The number most agencies leave out: repeat value

Almost every SEO business case is built on the first transaction and stops there, which systematically undervalues the work for any business where customers return. A dental practice, an accountant, a garage, a supplier of consumables — in all of them, the visitor who arrived through search once is worth several times the first invoice, and the model should say so.

The adjustment is straightforward: instead of margin on the first purchase, use margin across the expected relationship. You need two numbers from the client — average repeat frequency and typical retention period — and both are usually available from their own records even when nobody has calculated them before. The conversation that produces those numbers is often more valuable than the model itself, because a surprising number of business owners have never worked out what a customer is worth to them.

Two cautions keep this honest. Use conservative figures, because repeat value is the easiest place in the model to quietly inflate the answer, and a finance director will test that assumption first. And state clearly that repeat business depends on service quality, not on search — you delivered the first visit, the business earned the rest. That distinction costs you nothing in the model and buys considerable credibility in the room.

Where this matters most is in comparing channels. On first-transaction maths, paid search and organic often look similar for a local service business. On relationship maths, a channel that delivers customers at a lower marginal cost over time pulls clearly ahead — and that is precisely the argument organic search should be making.

The version of this that actually works

The business case that survives is short, conservative and explicit about its assumptions. One page with the four model inputs, three scenarios with named conditions, a clear statement of what is being promised — a scope of work and a range of outcomes — and an equally clear statement of what is not being promised, which is a specific position for a specific term on a specific date.

Everything else in a proposal is decoration. The keyword tables, the audit summaries, the competitor screenshots — none of them answer the question the person signing actually has, which is whether this is a better use of the money than the alternative uses in front of them.

If you have never built the calculation for an existing client, it is worth an hour. Open the dashboard, isolate the non-branded commercial group, and work out what the traffic it currently earns would have cost in paid search. That single number usually reframes the next renewal conversation entirely.