This article walks through a number of economic aspects of planning marketing budgets.
Among other things, you will learn:
- How to set an advertising budget?
- How to measure marketing results?
- Where is the optimal level of ad spend?
- Can you run campaigns without a budget?
Most of the examples and concepts refer to online advertising, but many of them generalize to all advertising activities.
What’s your budget?
A recurring theme in conversations between businesses and advertising agencies is the size of the advertising budget. The agency asked to prepare a proposal will usually ask about the budget.
The advertiser, in turn, often expects the agency to propose the amount of spend needed to achieve a given goal — or doesn’t want to reveal their budget, fearing the agency will use that knowledge solely to increase its own income, spending the client’s money lavishly.
This divergence usually stems from not understanding the other side’s situation. While large corporations typically operate on pre-set advertising budgets, in small and medium-sized businesses advertising budgets are allocated far more dynamically.
Uncertainty about the budget is especially common for companies just starting out, with no previous advertising experience, as well as when expanding into new markets or running a campaign in an advertising medium the company has never dealt with before.
The size of the advertising budget is a recurring theme between the advertiser and the agency.
Management expects the agency to support the decision based on the agency’s experience, overlooking the fact that the agency doesn’t know all the market conditions or the specifics and competitive position of the given product and company — and that directly transplanting campaign-effectiveness experience, even from very similar products, is very risky.
To present a proposal, the agency needs to know the budget framework. Proposing TV advertising to a client whose budget covers just a few airings of a spot misses the point. Agencies, in turn, often forget that by providing certain data they can help the client decide how big the budget should be.
The business plan
The classic method of setting advertising budgets is to allocate a fixed percentage of the company’s projected revenue to marketing. Naturally, there is no universal answer to whether it should be 1%, 5% or perhaps 40% of revenue.
The first, obvious factor is your margin. For high-margin products it is usually optimal to allocate a higher share of revenue to advertising.
Other relevant factors include the profitability of your advertising, the length of the period between a potential buyer’s first interaction with an ad and the purchase, the company’s investment horizon, and the scale of the planned expansion.
Still, based on financial projections we can establish a certain budget framework. What part of the sales margin are we able to allocate to marketing? Will it be 10%, 20%, maybe half? It is usually inadvisable for ad spend to lead to operating losses, so the margin tends to be an important boundary.
The budget must stay in reasonable proportion to your margin and other operating costs, and must be covered by the company’s financial capacity.
If, however, accepting an initial deficit, we plan a substantial investment in brand awareness through an aggressive advertising campaign whose effects will be felt for years to come, then the operating losses associated with that investment will have to be recouped within an acceptable time horizon — and above all, they must be covered by the funds available to finance such an investment, so that it cannot lead to insolvency.
On the other hand, if the advertising budget is a negligible fraction of your profits, your marketing is most likely too defensive. In a competitive market economy, the advertising market is highly efficient. Market prices of ad inventory reflect the benefits advertisers derive from it, and opportunities offering extraordinary returns are extremely rare and usually short-lived.
That’s why the advertising budget must stay in reasonable proportion to the margin and other operating costs, and match the company’s financial capacity. Business-plan data will let us establish at least the order of magnitude of the advertising budget. Such information will certainly be very useful to the agency, which can then prepare several sensible budget options.
The campaign goal
A clearly defined campaign goal makes budget estimation much easier. Suppose the goal is to sell 200 stays at a holiday resort worth €3,000 each. The margin is €1,500 per stay. Based on this information alone we can start estimating the budget. We’d like the marketing cost not to exceed 20% of the margin, i.e. we don’t want to spend more than €300 to sell one stay, which means the budget should not exceed 200 × €300 = €60,000 — though of course we’d like to spend as little as possible.
We plan to run the campaign in Google search. We know that among people who search for the relevant keywords and land on our offer page, one in 20 to 50 submits an inquiry (visit-to-inquiry conversion of 2 to 5%). We also assume that one in 5 to 10 inquiries leads to a sale (inquiry-to-sale conversion of 10 to 20%). We can therefore estimate the number of site visits we need to sell 200 stays. It will be at least:
20 × 5 × 200 = 20,000 visits
and at most:
50 × 10 × 200 = 100,000 visits
The agency tells us that the cost of a single visit falls between €1 and €2, which means a campaign that achieves our sales goal will cost between €20,000 and €200,000.
The worst-case scenario (€200,000 of spend) means that having sold 200 stays, we spent €1,000 to sell each one. That is still less than the margin and seems, in the worst case, still acceptable.
From this data we can clearly see that a budget of, say, €10,000 is not realistic, and we can hardly count on achieving the goal for that amount.
Taking advantage of the fact that in search PPC (pay-per-click) advertising the daily budget can be modified at any moment, we set the budget at, say, €50,000 to €100,000, accepting that based on the first performance data we will have to decide whether to increase or decrease it.
Reach and frequency
The goal of online advertising will not always be generating traffic and sales on the advertiser’s website. Very often clicks will be of secondary importance.
TV, press and radio ads don’t get clicked, yet nobody questions their effectiveness. In most cases, advertising FMCG brands have nothing special to offer on their websites and don’t sell online, since their products are sold through distribution networks.
If the campaign’s goal is to reinforce the brand image and the product’s competitive advantages, this is achieved by reaching the target group with a message contained in the ad creative itself, with no need to redirect anyone to a website.
We may also deal with a mixed goal. A consumer-electronics retailer wants to persuade us to buy a TV on promotion, but is also aware that the ad will reinforce the brand image among those who aren’t currently interested in buying — for example because they recently bought a TV.
In such cases we operate with the categories of reach and frequency. Reach tells us how many people saw the ad, and frequency — how many ad impressions there were per unique user. Knowing the size of the target group and the desired frequency, we can calculate the necessary number of ad impressions:
Impressions = Reach × Frequency
For example, a wedding dress manufacturer wants to run a Facebook campaign. The agency informs us that over the past 6 months, 31,000 Polish female Facebook users marked their engagement on their profile. We’d like to reach at least 25,000 of them (reach), showing them our ad 3 to 5 times on average (frequency).
This means we’ll need to buy between 75,000 and 125,000 ad impressions. The agency also estimates that the cost of 1,000 impressions (CPM) on Facebook is currently around €4, which puts the potential budget between €300 and €500. Even if we assume a frequency of 20, the budget won’t exceed €2,000.
If our budget — resulting from our financial capacity or an assumed percentage of turnover — is €40,000, allocating all of it to a one-month Facebook campaign seems pointless. We should rather think about extending the campaign, broadening the target group, or looking for another advertising medium, e.g. Google search, price comparison sites, or wedding portals.
Not all reach is created equal
When comparing impression prices (and thus the cost of reach) across media, we must remember that these are very often incomparable measures. The mere fact of an impression means little. What matters is the probability that the displayed ad was noticed by the user, whether and for how long their eyes rested on it, and to what extent the advertising message actually got through.
Look at the illustration below. The same ad (a sponsored post) on Facebook can be shown to a user as a small section in the right column, or in the news feed, between posts published by the user’s friends. It is obvious that the right-column ad has incomparably less impact.
Does this mean you should avoid right-column ads and focus only on the news feed? No. It turns out that the cost of a thousand right-column impressions is several times lower. They will, however, require higher frequency to achieve reach comparable to news feed ads.
The impact of a large graphic creative displayed at the top of a web page will be much greater than a small text ad tucked into the bottom-right corner. This will most likely be reflected in a severalfold difference in CPM.
The higher CPM of YouTube video ads reflects the fact that the spot plays directly before the video the user actually wants to watch, so it is highly likely that the user’s eyes are focused on the ad.
Naturally, the value of impressions will depend on the target group, and will usually reflect that group’s purchasing power. The cost of impressions shown to teenagers will be entirely different from those shown to mature audiences interested in finance or luxury goods. Context also matters — impressions on prestigious sites thematically related to the ad will potentially be worth more.
For this reason, applying a single measure to all ad impressions is pointless. There is no universal answer to the question of how much an ad impression is worth.
Why PPC (pay per click)?
One of the most popular advertising billing models is pay per click (PPC). Its advantage for the advertiser is that it prices advertising according to a measurable user interaction with the ad (the click), which can additionally be verified in the advertiser’s own website analytics.
Even if clicks are not the campaign’s primary goal, the clickthrough rate (CTR) — the ratio of clicks to impressions — lets us roughly compare the actual impact of different creatives in a given placement. Large creatives in prominent positions will have higher clickability, and small creatives placed out of the user’s sight will be clicked very rarely. At the same CPC, differences in the cost of displaying these ads will approximately correspond to differences in their impact.
Remember, though, that clickability also depends on how much the creative itself invites a click. With similar exposure, an ad saying “Our new candy bar will give you energy” will most likely be clicked far less often than an ad with the slogan “Eat our bars and win a car” and a “Join the promotion” call-to-action button. So when using CTR as a measure of actual user reach, the content of the creative must also be taken into account.
Most advertising systems are now based on real-time auctions. Advertisers compete for available ad inventory by stating the highest price they are willing to pay. The price set in the auction is the minimum price needed to outbid the competitor’s maximum. In many ways the advertising market resembles financial markets, where the price is set by the interplay of supply and demand. Individual placements are priced by their buyers based on effectiveness. As a result, the online advertising market is largely efficient and exceptional bargains are rare.
Just like in financial markets, in online advertising “there are no free lunches”. One typical misunderstanding of the PPC model is the hope of creating ads that won’t be clicked, thereby earning free impressions. A similar misconception is the fear that ad clicks will expose us to losses.
One could say that publishers (Google, Facebook, the sites where ads appear) don’t care about our clicks. Their goal is to maximize revenue from ad space. For ads that are clicked less often, the publisher will expect higher per-click rates, while creatives that get clicked frequently will buy clicks ever more cheaply. And although in reality systems like Google or Facebook use not only CTR but also other ad quality criteria when converting CPC into an effective cost of impressions, what ultimately counts is the effective use of ad space — which nobody wants to give away for free. The PPC model merely makes it easier for advertisers to price the value of advertising.
Measuring advertising effectiveness
Thanks to website analytics tools such as Google Analytics (available for free), advertisers can analyze in detail the effectiveness of advertising related to their website. We can measure the performance of individual traffic sources and campaigns, analyzing not only the number of visits but also their quality — through measures such as time on site, number of pages viewed, or conversions, i.e. completing a specific action (submitting an inquiry, making a purchase, downloading materials, or visiting the contact page).
The analysis can run along many dimensions. We can and should segment statistics by traffic source, type of creative, device (desktop, tablet, phone), geography, time of day — to name just a few of the most important.
For online stores, we can precisely measure how many visits and how much revenue a campaign generated, and which products it sold.
Online statistics aren’t everything
The ability to precisely measure and analyze traffic generated by online advertising is a blessing — but often also a curse — of online marketing.
Thanks to web analytics, we no longer have to take the press publisher’s word that 100,000 copies sold are read by 200,000 people who happen to be a well-profiled target group. We see the traffic and transactions in our reports. This, however, leads many advertisers to conclude that what Google Analytics shows is everything their online campaigns achieved.
Remember — that’s not the case! First of all, there are impressions without clicks, which certainly contribute — to a greater or lesser extent — to raising brand awareness and to later direct visits.
When analyzing transactions attributed to a campaign, they should be considered in the context of the total value of acquiring a customer. In many industries, winning a customer means more than a single transaction.
Depending on the product and how it is sold, beyond the direct revenue from online conversions an online campaign can bring other benefits, such as offline sales (e.g. by phone or physical store visits), orders originating from email inquiries, as well as increased brand awareness and consequently future conversions from direct sources, sales to returning customers, or sales resulting from customers recommending the product to others (Lifetime Value).
Lifetime Value should be estimated within a time horizon acceptable to the advertiser, which may depend in particular on the availability of funds for long-term investments.
As you can see, the revenue from the current transaction resulting from a conversion is merely the tip of the iceberg.
Regardless of how difficult these additional benefits are to measure, it’s worth estimating them and including them in the conversion revenue used to calculate campaign ROI. Without accounting for Lifetime Value in conversion value, our assessment of campaign value will be defensive — it can limit expansion and lead to being pushed out of the market by competitors.
Managing ad spend without a budget
Imagine a transport company that set rigid fuel budgets in its financial projections. At some point the company receives a large order for very profitable haulage. Unfortunately, it is forced to turn it down, because it has already exhausted its fuel budget for the period and cannot fill up its trucks.
The example is, of course, absurd. Even if such a budget formally exists, any sane manager will decide to allocate additional funds for fuel, since the investment in fuel pays back almost immediately in service revenue — any other decision would expose the company to lost profits.
Similarly with advertising: a rigid advertising budget will usually be suboptimal. There is no way to predict in advance how effective a campaign will be and where the optimal spend level lies. A budget set a priori will always be either too small or too large.
That’s why the ability to modify and shift budgets is an opportunity for extra efficiency. Here too, communication between advertiser and agency is key, and decisions should be based on business goals.
Picture a travel agency selling flight tickets. At the peak of the season, the number of flight-related Google searches soars, and with it the potential budgets that can be spent. The season also usually brings higher conversion rates — a larger share of visitors end up buying. Restricting the budget would mean giving up a large volume of profitable ticket sales, while allocating extra budget brings extra profit.
When the season ends and customer interest drops, campaign profitability usually drops too. Very often in such a period it pays to cut ad spend and accept lower sales. Force-spending an oversized budget at that time may not translate into profit. So it’s worth increasing budgets in periods of higher effectiveness, and cutting them when advertising profitability falls.
As a side note, it’s worth mentioning the self-regulating mechanism of search PPC. In periods of lower user interest (e.g. off-season or at night), the number of searches drops as well. Few searches means few ad impressions and clicks — and therefore lower costs.
So ads shouldn’t be switched off; at most, adjust the bids if conversion rates differ in these periods. Switching off search ads in periods of lower interest makes, as a rule, about as much sense as switching off the light in the fridge.
Switching off search ads at night makes as much sense as switching off the light in the fridge.
An airline may see it somewhat differently. In season, when flights sell like hot cakes and all seats are sold out, investing in direct-response advertising (ticket purchases) may be pointless. Off-season, when planes often fly nearly empty, running even a less profitable campaign may be justified in terms of opportunity cost.
A ticket reseller’s perspective will differ again, being able to respond more flexibly to demand shifts and not carrying such high fixed costs.
This is precisely why very many e-commerce companies run campaigns without an advertising budget. The budget is treated as unlimited for as long as the campaign meets its effectiveness targets — cost of customer acquisition or return on investment (ROI).
The campaign manager’s aim is to maximize sales within that target. For example, a marketer may spend practically any amount on a campaign as long as it delivers at least 50% profitability.
By measuring campaign effectiveness on an ongoing basis, the advertising budget is adjusted to the campaign’s current capacity to generate profitable sales.
Marginal cost per conversion
When a campaign generates sales at the assumed profitability, the expectation often arises to increase ad spend and thereby generate higher revenue — and thus profit. In such a situation, campaign profitability must be considered in terms of marginal effects.
Imagine an advertiser selling products online. On average, each transaction earns them €200. So they agree with the campaign manager on a target cost per conversion of €150.
The campaign steadily generates 1,000 conversions a month on a spend of €120,000, i.e. €120 per conversion. One day a decision is made to increase the campaign’s intensity. As a result, the following month brings 1,300 conversions on a spend of €195,000, i.e. €150 per conversion. Everything seems fine: sales are growing and the cost per conversion is under control.
Is it, though? Let’s calculate the cost of that sales growth. The campaign change brought 300 additional conversions at a cost of €75,000 more. This means those additional conversions cost €250 each — so we lost €50 on every one of them.
| Conversions | Campaign cost | Campaign revenue | Campaign profit | ROI | Cost per conversion | |
|---|---|---|---|---|---|---|
| Optimal campaign | 1,000 | €120,000 | €200,000 | €80,000 | 66.7% | €120 |
| Aggressive campaign | 1,300 | €195,000 | €260,000 | €65,000 | 33.3% | €150 |
| Change | +300 | +€75,000 | +€60,000 | -€15,000 | -20% | €250 |
This example perfectly illustrates how misleading it can be to set a campaign goal in the form of a fixed cost per conversion.
To manage cost per conversion properly, you need to work with the concept of the marginal cost per conversion — how much it costs to acquire the next additional conversion.
The more you buy, the more it costs
In online campaigns, marginal cost usually rises as campaign intensity and scale grow. As we already know, a large part of the ad inventory market is priced in an auction in which participants can freely change the prices they are willing to offer. Such an auction means big players don’t pay less. The unit price rises with campaign intensity.
Take Google search advertising. To acquire a small amount of traffic, it’s enough to show your ad at lower positions, paying lower per-click rates. To buy more traffic, you’ll have to offer higher bids to outbid competitors and move up the positions.
By raising the offered CPC we also increase traffic, though as the bid grows, the traffic gains usually get weaker and weaker, and at some point, despite raising bids, we can buy practically no additional traffic. We say that the marginal cost of acquiring a user rises to infinity.
Below is an example plot of the function relating the number of clicks to the cost per click, Clicks(CPC):
This principle can generally be extended to media planning across the entire web. The more traffic we want to acquire and the more we expand the campaign’s reach, the higher the marginal cost of the next user we must expect.
For this reason, campaign intensity can be increased for as long as the marginal cost of acquiring a customer stays below the amount that customer earns us — that is, until we start losing money on the next user acquired.
Put differently, the marginal profit from the campaign should equal zero. This strategy truly maximizes campaign profit. Increasing ad budgets makes sense for as long as the marginal (not total) profit is greater than zero.
When to spend more?
Return on investment (ROI) — a measure borrowed from methods used to assess the profitability of financial investments — is not suitable as a profitability target for online campaigns. The campaign’s goal should be maximizing profit, not the rate of return, and decisions to change bids should be made based on marginal profit analysis.
Is increasing sales by 10% while raising the cost per conversion by 5% worthwhile for us? Where is the line beyond which expanding the campaign will increase sales but worsen the bottom line? Questions like these can only be answered by analyzing the marginal profit from conversions.
It turns out that the relationship defining the optimal spend level can be described with a simple formula: ROI > 1/E, where ROI is the return on ad spend and E is the price elasticity (you’ll find the derivation of this formula in this article).
Overinvestment and underinvestment zones
The cost per click rising with campaign expansion and the falling price elasticity mean that the marginal profit per conversion decreases, at some point turning negative.
As we already know, the point where marginal profit is 0 and ROI equals 1/E is the point of maximum total profit and marks the optimal level of investment in advertising. For as long as ROI > 1/E, the campaign is underinvested and increasing spend will increase profit.
Beyond the optimal level ROI = 1/E we enter the zone where ROI < 1/E. Sales keep growing, we’re still making a profit, but it keeps shrinking. At some point ad spend exceeds income, producing an operating loss.
Unfortunately, it is often only when operating losses appear that advertisers notice they have overinvested in advertising. Yet overinvestment starts much earlier — the moment we pass the point of maximum profit and ROI falls below 1/E.
A marginal analysis example
Suppose that in our Google Ads campaign, using the A/B experiment feature, we test new campaign settings in which a higher bid gets us a higher position, gaining more clicks — though at a higher cost per click (€1.67 vs €1.41), which gives a cost per conversion of €167. Our revenue per conversion is €200, so the campaign is profitable.
But are these settings actually good for us? Calculating the cost of acquiring the additional 323 conversions, we see it amounts to €199.60 — close to the break-even threshold, but still below it.
| Max CPC | Impressions | Avg. position | CTR | Clicks | Avg. CPC | Cost | Conversions | Cost per conversion | ROI | 1/E | |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Control | €1.80 | 500,000 | 3.1 | 8.1% | 40,500 | €1.41 | €57,105 | 405 | €141.00 | 41.8% | 0.21 |
| Experiment | €2.10 | 650,000 | 2.7 | 11.2% | 72,800 | €1.67 | €121,576 | 728 | €167.00 | 19.7% | |
| Total | 1,150,000 | 2.9 | 9.9% | 113,300 | €1.58 | €178,681 | 1,133 | €157.71 | |||
| Change | +16.6% | +150,000 | +32,300 | +€64,471 | +323 | €199.60 |
This means the experimental bid will deliver higher profit. So we raise our bids to the experimental level, and in the next experiment we test what happens to the cost per conversion when bids are raised further:
We can see that the marginal cost per conversion has now exceeded €200, reaching €223.96. Such an aggressive campaign is no longer profitable, even though the average cost per conversion is still below €200.
| Max CPC | Impressions | Avg. position | CTR | Clicks | Avg. CPC | Cost | Conversions | Cost per conversion | ROI | 1/E | |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Control | €2.10 | 650,000 | 2.7 | 11.2% | 72,800 | €1.67 | €121,576 | 728 | €167.00 | 19.7% | 0.25 |
| Experiment | €2.20 | 655,000 | 2.5 | 13.2% | 86,460 | €1.76 | €152,170 | 865 | €176.00 | 13.6% | |
| Total | 1,305,000 | 2.6 | 12.2% | 159,260 | €1.72 | €273,746 | 1,593 | €171.89 | |||
| Change | +4.7% | +5,000 | +13,660 | +€30,594 | +137 | €223.96 |
Using the ROI > 1/E formula, we can significantly speed up decision-making and avoid costly experiments. At the first bid increase, the price elasticity was 4.8. This means 1/E = 0.21 — less than the current ROI (41.8%), so bids could be raised. At the second bid increase, elasticity fell to 4.0, i.e. 1/E = 0.25 — more than the current ROI (19.7%), which means raising bids no longer made sense at that point.
What is the right ROI?
We now know that to determine the optimal ROI — the perfect compromise between sales volume and margin, leading to maximum profit — two measures must be compared: the current ROI and the current price elasticity. It is price elasticity alone that determines the appropriate profitability level.
Price elasticity is not constant across a campaign. Where bids are already high, elasticity is usually lower. This means that applying the same target ROI to all campaign segments is a mistake.
A typical effect of the single-ROI strategy in search is overinvesting in the best-performing campaigns while marginalizing those with somewhat weaker performance — so we end up reaching almost exclusively for the lowest-hanging fruit.
Compared with such a strategy (a fixed ROI for the whole campaign), optimization will — paradoxically — consist in lowering bids for the best-converting campaigns and raising them for the weaker-converting ones.
Reducing investment in areas of low price elasticity and moving it to areas where price elasticity is high will increase sales volume within the same budget.
So what is our budget?
We come back to the question posed at the beginning. There is no universal formula for determining the optimal advertising budget, and covering all aspects of optimizing marketing spend goes far beyond the scope of this article.
When planning an advertising budget, it’s worth combining the company’s financial data and knowledge of its business goals with the know-how of the agency, which can estimate the size of the target group and the approximate costs of reaching it.
The advertising budget should stay in reasonable proportion to the company’s other costs and account for the actual cost of the advertising media, which must fit both the budget framework and the campaign goals. A clearly defined goal makes estimating the required spend easier.
In today’s reality, especially in online marketing, the intensity of campaigns and ad spend can be modified almost in real time and adjusted to current effectiveness. The high measurability of results lets us treat advertising costs not as a cost, but as an investment.
As a result, the optimal solution is very often to abandon the rigid advertising budget and allow it to be modified. This helps avoid missed opportunities on the one hand, and overinvestment in advertising on the other.
Budget modification decisions should be based on marginal measures, comparing the campaign’s current profitability with its price elasticity. There is no single right ROI for all advertising activities. Where price elasticity is high, we can accept a lower ROI, since the reduced profitability is compensated by substantial gains in sales volume.
