What an Rp 65 Million Search Retainer Actually Funds

Search Agency's Growth tier starts at Rp 65 million a month. That money buys an agreed amount of working capacity, a fixed set of workstreams and a reporting rhythm. It does not buy a ranking, a citation in an AI answer, or a guaranteed number of leads, and any agency selling those is selling something it cannot control.

This article explains what sits behind the number. It is written by Search Agency about its own price, so read it as a seller describing its offer. Everything here comes from our published pricing, methodology, measurement and audit pages as they stood on 2 November 2026. The signed statement of work for any engagement governs the details, and some details below are marked as things the statement of work has to settle.

What the price is buying

The pricing page lists four tiers. A free Benchmark Audit sits at the bottom, Foundation starts at Rp 35 million a month, Growth starts at Rp 65 million, and Enterprise starts at Rp 150 million. Every tier bundles technical SEO, AEO and GEO into one engagement. AEO (answer engine optimization) is the work of getting your pages cited in answers such as Google's AI Overviews. GEO (generative engine optimization) is the work of getting your brand named and recommended by assistants such as ChatGPT, Gemini and Perplexity.

The Growth tier is described in four lines on the page.

• The full Benchmark, Engineer, Earn and Measure cycle
• Hub-and-spoke content and answer-first rewrites
• A digital PR and off-page authority program
• Coverage of three to five AI engines, quarterly re-benchmarks and a dedicated pod

The word "from" matters. The page says prices are starting points, that every engagement is scoped from the audit findings, and that the number you sign reflects your market, your competitors and the work required. Rp 65 million is the floor for this scope. A larger site, a harder category or more markets will price above it.

The useful way to read a retainer is as operating capacity. You are paying for a team's time, pointed at an agreed plan, with outputs you can inspect each month. What that capacity achieves depends on the plan, on your side's ability to ship the work, and on search engines and AI models that nobody controls.

Who does the work

The pricing page says Growth includes a dedicated pod. It does not publish how many people are in it, their seniority, or how their hours divide between accounts, and we will not invent those details here. Your statement of work should name the roles, the person accountable for your account, and how much of each role's time is committed to you.

What we can describe is the work that has to be covered, because our methodology sets it out as eight stages, and each one needs a skill set.

Two plain-language notes help here. An entity is the way search engines and AI models identify your brand as one specific organization, and schema is code on your pages that states facts about that organization in a format machines read. Stage seven on our methodology page, which we call Total Graph Authority, is the point where your own pages, third-party sources and social signals describe the brand the same way, and stages two to six build toward it.

Capacity is finite, so these workstreams compete for the same hours. A month that needs heavy technical work after a site migration leaves less time for content. A month with a large PR push leaves less for rewrites. The monthly readout is where that trade-off gets decided in the open, with you in the room.

What gets delivered

Our methodology page states that every stage ships something you can hold us to. Those outputs are the evidence that the retainer is being worked.

• A scored baseline from the benchmark, which becomes the set of numbers you judge us on later
• A prioritized defect list from the technical stage, with our team working alongside your developers while they clear it
• An entity map and the structured markup deployed against it
• A content map that you see before anyone writes a word, followed by the content and rewrites it calls for
• A list of named third-party targets set at the start, with reporting on what landed
• A live dashboard that starts at stage one and keeps running
• A monthly readout of thirty minutes that ends in decisions about next month
• A quarterly re-benchmark against the same locked prompt set

In practice these arrive as working documents. The defect list becomes tickets your developers can pick up. The content map becomes briefs, then drafts for your approval, then published pages. The authority work becomes an outreach record showing who was approached and what came of it. The readout becomes a short decision note that says what moved, what probably caused it and what changes next.

The public pages do not fix a monthly content volume or a number of placements for Growth. Those quantities belong in the statement of work, set from the audit findings, and you should expect to see them written down before you sign.

How the first months differ

The first quarter is mostly diagnosis and repair. The methodology page puts the first benchmark at about two weeks and says the release path is scoped in week one. Our SEO service page describes a discovery stage of two to three weeks and a 90-day roadmap that is approved before any execution starts.

The weekly timings in this table are an illustration built from the stated milestones. Your own roadmap sets the real dates.

The two phases move at different speeds. The methodology page says technical and entity work ships in weeks and registers with crawlers quickly. Third-party validation is the slowest stage, because it runs on the timelines of editors, reviewers and industry bodies who owe you nothing. That is why the targets are named at the start, so that progress is visible while it is slow.

From the second quarter the work becomes steadier. The prompt set stays frozen so that the trend line means something. On our measurement page we say we revisit the weighting of prompt types each quarter and leave the prompts themselves alone.

What the client must provide

A retainer only works when both sides do their part. Our methodology page names three things that sit with you, and they decide more outcomes than most buyers expect.

The first is a release path. We write the schema, the redirects and the fixes, and someone on your side ships them. The methodology page says more engagements stall here than anywhere else. If your developers have a long queue or a quarterly release cycle, plan for that before signing.

The second is access to a subject-matter expert. Content that models cite needs someone who knows the category, and we ask for about an hour a fortnight with your product or specialist lead. Without that time the content tends to be generic, and generic content is easy for a model to skip.

The third is someone who can decide. The monthly readout ends in a decision about next month. When the only attendee has to take every decision back to a committee, our page notes that the loop stretches from four weeks to twelve, and the trend line stops meaning much.

You also need working analytics. Broken GA4 setups and half-fired tags are a separate job. Our analytics engineering team, Awesonatics, handles that work, and we will tell you when your tracking cannot support the reporting you want.

What is outside scope

Some things are excluded because they cannot be bought, and some are excluded because they are separate work. Our public pages state the following.

Several cost items are not settled by the public pages, so they have to be settled in your statement of work. These include tool subscriptions, developer hours on your side, any paid media or placement fees, translation, photography or video production, taxes, and how change requests are priced. Ask for each of these in writing before you compare our quote with anyone else's, because two quotes that treat them differently are not comparable.

What changes the price

The pricing page says the signed number reflects your market, your competitors and the work required. Our article on what really sets the cost of SEO sets out the drivers in more detail. For a Growth engagement they tend to show up in five ways.

• The number of markets and languages, because each one needs its own prompt set, benchmark and content
• The size and complexity of the site or portfolio, since a large catalogue or several brands means more templates to fix and more entities to resolve
• Technical debt, meaning the backlog of crawl, rendering and duplication problems that must be repaired before growth work pays off
• How competitive the category is, because stronger rivals need more authority work to move against
• Delivery intensity, meaning how much you want done in parallel, which needs more people working at the same time

Speed is the driver buyers misread most often. Paying for more parallel work can shorten a roadmap, but it cannot make crawlers, editors or AI models move faster. Past a point, a higher fee buys sequencing and focus, and very little extra pace. For the range of prices across the Indonesian market, see our breakdown of what an SEO and AI search campaign costs in Indonesia.

When this tier is wrong

Growth suits a brand that is competing for a category, can ship fixes, can give expert time, and can wait for authority work to compound. When those conditions are missing, a different route usually gives better value, and several of them cost less or nothing.

A smaller engagement fits when the main need is foundations. Foundation starts at Rp 35 million a month and covers technical SEO remediation, schema, entity foundations, tracking across three to five AI platforms and a monthly dashboard, without the content build and PR program. If you already pay another agency for SEO and only lack AI visibility, the pricing page lists a GEO Sprint, a 90-day add-on priced at 20 to 40 percent of your existing retainer that keeps your current vendor.

An internal hire fits when you already have the people. Our methodology page says a technical SEO who can get schema and crawl fixes deployed can run stages two and three in-house, and a content team that holds a hub-and-spoke map for months can run stage four. Stages five and six, the authority and social proof work, are the hardest to staff internally because they depend on relationships.

A project fits when you need a decision more than a program. The free Benchmark Audit is a fixed-scope deliverable with a 15-page report, a 30-minute walkthrough and a prioritized 90-day roadmap you can act on with us or in-house, with no obligation to continue. It is available to brands we could realistically take on, and every request is reviewed first.

Postponement fits when the basics are not ready. If nobody on your side can release fixes this quarter, if your tracking cannot tell you where enquiries come from, or if buyers in your category do not use AI tools in a way that shapes their choice, a retainer will spend money waiting. Our methodology page says that when a category has no meaningful AI answer volume, we will tell you the budget belongs elsewhere.

How performance is reviewed

Performance is reviewed in three layers, and each layer answers a different question.

Delivery obligations answer whether we did the agreed work. These are the stage outputs listed above, delivered on the agreed schedule, and they are fully within our control.

Leading indicators answer whether the work is registering. Our measurement page defines them. Mention rate is the share of tracked prompts where an answer names you. Citation share is the share of answers with citations that cite your domain. Recommendation rate is the share of mentions where the model actively puts you forward. Share of voice is your mentions against all brand mentions in your competitive set. Prompts run weekly across three to five assistants, three competitors are scored alongside you, and we publish the weights behind the composite visibility score (35, 30, 20 and 15 percent) so the number can be audited.

Commercial outcomes answer whether the business benefits. Clicks, leads and revenue from AI surfaces are tracked, but our own measurement page says this layer is correlational and is never presented as causal proof. Results are also judged across quarters, because one month is too short to separate the effect of the work from model noise.

The table below shows how commitment, dependency and uncertainty sit across the main areas of a Growth retainer.

If the numbers do not move, the methodology page commits us to show which stage is blocking and what it would take to clear it, including when the blocker sits on your side. Check whether the Growth scope fits your market and delivery capacity by comparing the tiers on our pricing page, or talk to us about your scope.