Others have reflected on the framing of the Aichi Targets (Wood 2011; Butchart et al. 2016), but we are the first to carry out a SMART‐based assessment of all 20 targets and to show evidence of associations between progress and the extent to which the Aichi Targets were perceived to meet certain criteria. We found significant positive relationships between progress and measurability, unambiguity, realism and scalability of the target elements, and some evidence of a positive relationship with specificity. There was little variation between target elements in the scores for ambition, and we found no significant relationship between ambition and progress or realism. Because all Aichi Targets are explicitly time bound, this criterion was not assessed, but is an important characteristic when setting targets and should continue to be incorporated into any new target framework. Our results should be taken as indicative because they were based on the responses of 49 survey participants, most from the UN Regional Group's Western European and Others Group (Supporting Information). Although not feasible in this study, a longer time frame may have allowed a larger sample size with a more balanced geographical distribution and diverse set of opinions.

The mechanisms underpinning the relationship between target framing and progress were not explored here, but warrant further investigation. It is plausible that governments find it easier to identify and implement actions and policies if a target element is more specific and measurable and is unambiguously worded. Elements that are vague and open to interpretation may be more difficult to translate into specific actions or policies. Given that nearly all actions to address targets are applied nationally or locally, the association between progress and scalability of target elements is unsurprising. Similarly, the strong association between progress and target element realism is unsurprising: commitments that are unrealistic are less likely to be achieved.

Many factors will influence progress toward biodiversity targets. The midterm assessment by Tittensor et al. (2014) reported greater progress toward targets that covered responses to biodiversity loss than those targeting pressures on, states of, or benefits from biodiversity. This may be due to responses being easier to implement and measure than states and pressures and time lags between response implementation and consequent positive effects on biodiversity or the implemented responses being insufficient or inappropriate (Tittensor et al. 2014). The response targets (goal C) generally scored higher in our assessment than other targets, whereas targets focused on mainstreaming (goal A) and reducing pressures (goal B) scored lower. This indicates that targets related to the state of biodiversity and conservation responses to biodiversity loss were considered to more closely align with the SMART criteria used in our assessment than those targets relating to raising awareness, sustainable use, and pressure reduction, which strengthens the argument that responses are easier to action and measure. We found a large bias in research effort toward Target 11 (with a focus on protected areas in particular, a conservation response), which, in terms of terrestrial areal coverage, is on track to be achieved by 2020 (Secretariat of the CBD 2014). The area‐based elements of Target 11 had high SMART scores, whereas 4 qualitative elements (which call for “areas of particular importance for biodiversity and ecosystem services” to be conserved through “effectively and equitable managed, ecologically representative, and well‐connected systems of protected areas”) were perceived to be substantially less SMART (Supporting Information) and are unlikely to be achieved by 2020 (Secretariat of the CBD 2014). The mismatch in progress toward the more SMART quantitative elements and less SMART qualitative elements of Target 11 suggests the effectiveness of protected areas for conserving global biodiversity is currently limited by inappropriate placement and ineffective management. Protected areas often fail to capture ranges of threatened species (Beresford et al. 2011; Venter et al. 2014; Butchart et al. 2015) or are located in remote, mountainous areas that are under little pressure (Joppa & Pfaff 2011) and fail to halt the loss of natural land cover within their boundaries (Geldmann et al. 2013). Topography (elevation and slope) can be as effective at minimizing land cover change as designation of protected areas (Beresford et al. 2017).

The Aichi Targets have been described as more holistic and detailed, SMARTer, and better at addressing drivers of biodiversity loss than the CBD 2010 Biodiversity Target and other international targets (Harrop & Pritchard 2011; Collen et al. 2013; Mace et al. 2013; Harrop 2014; Rees et al. 2018). However, they have also been criticized for ambiguities, unquantified objectives, a lack of baselines and indicators, and insufficient objectives (Jørgensen 2013; Mace et al. 2013; Harris et al. 2014; Kadoya et al. 2014; Polak et al. 2015; Butchart et al. 2016; Heywood 2016; O'Leary et al. 2016; Beresford et al. 2017). Several studies suggest improvements that could be made when framing future biodiversity targets. These include ensuring that targets are SMART (Wood 2011; this paper) or meet other criteria, such as comprehensive, understandable, and enabling (Bridgewater 2011). Future frameworks could also consider employing a smaller number of more focused headline targets that are specific, quantified, simple, succinct, and unambiguous and that are accompanied by more specific subsidiary targets (Butchart et al. 2016). However, adhering to a SMART or similar framework during target setting, particularly around contentious issues, may be hindered unless sufficient focus is paid to the process of consensus building between stakeholders with diverse and often conflicting views (Maxwell et al. 2015). Targets should also reflect levels of conservation actions that are sufficient to deliver the protection and restoration of biodiversity, based on the best available scientific evidence (Di Marco et al. 2016). For example, to aim sufficiently high to halt the loss of biodiversity, there have been calls to increase the values of some quantitative elements, such as those of Target 11 (Locke 2013; O'Leary et al. 2016; Dinerstein et al. 2017; Krueck et al. 2017).

We recommend that any new or revised targets established under a post‐2020 global biodiversity framework should be clearly and unambiguously worded so that the intent and necessary action or actions are apparent; well defined with explicit deliverables and include quantifiable elements where appropriate so that progress toward the target can be measured; scalable so that actions can be implemented at global, national, and regional scales; and realistic so as to encourage action without being perceived as unachievable. See Supporting Information for an example with Target 7 (purely a demonstration of the application of a SMART framework). Although we cannot draw direct conclusions regarding ambition from our analyses, others have argued that targets need to aim sufficiently high to have positive outcomes for biodiversity and inspire action, but not so high that they become unrealistic to the point of reducing accountability and discouraging progress (Maxwell et al. 2015; Hagerman & Pelai 2016; Mace et al. 2018). We recognize that the target‐setting process is influenced by political practicalities and conflict resolution between diverse stakeholders (Larsen et al. 2015; Maxwell et al. 2015; Rodriguez‐Rodriguez et al. 2016). Nonetheless, Di Marco et al. (2016) called for scientists to be more engaged during target setting, specifically by providing policy makers with “direct evidence of how alternative formulations of targets […] can lead to improved biodiversity outcomes.” As negotiations around the post‐2020 global biodiversity framework proceed, we hope our analysis encourages decision makers to adhere to a SMART‐based framework as closely as possible to provide a strong basis for targets that are consistently interpreted, effectively applied, and adequate for the conservation, restoration, and sustainable use of biodiversity and ecosystem services.